So after yesterday's New Coke post, it's back to football around here. And while I wish the Sixers the best of luck with their new pick, man, you guys have your work cut out for yourselves when it comes to making people care.
Continuing the semi-theme of looking around the division, I spent some time looking at the Cowboys offense last night. I was curious about a few things, like why the offense wasn't as good last year, how Roy Williams should fare this year, and just generally how has the shape of it evolved since Romo took over two and a half seasons ago.
This isn't going to be one of those argument posts, where I make a case for one reading of the data. Instead, it's more descriptive, with some new questions at the end.
The first player I looked at was Jason Witten. For my money, he's the best skill-position player Dallas has, but I do think tight end is a position with diminishing marginal pass-catching returns. That is, you want a great tight end to make situational plays and help open up the rest of the offense, but if you go to him too many times, you may be missing opportunities on the outside for bigger plays. (TO probably wouldn't describe the situation the same way, but he'd no doubt agree.)
The traditional stats give some support to this position:
Almost 100 receptions for a tight end is a very big number. Witten's yard/catch number has been remarkably stable, but has their been a per-play downside to "forcing" him the ball? For that, we turn to FO's stats:
Uh, no. It hasn't been a problem. His DVOA dropped a bit last year, but you're still looking at remarkably consistent production, whether you throw him the ball 91, 120 or 141 times.
What about Marion Barber? That's a guy whose role really changed last year. He went from being the "closer" to the "man," and it doesn't seem to have been a good shift:
And unlike Witten, Barber's FO stats don't bail him out:
That's a big time regression right there. Of course, when we're talking about running backs having issues, the first thing to check is always the offensive line. The problem for Barber is that the rest of the backs didn't share his issues:
Yes, Barber had the toe injury late last year, but check out his game logs. They don't tell the story of a guy who was playing great until he got hurt.
Especially given how well the other two guys ran the ball last year, and how dangerous Felix Jones can be when he has the ball in his hands, I don't see how Dallas doesn't shift Barber back to his old role as a late-game punisher. Sure, let him start because he's earned it, but give the other guys the mid-field and early carries, and bring in Barber when it's late or you're down by the goal line.
A new coaching staff would probably have no problem doing that. Fortunately, Wade Phillips is still around.
Here's another chart that suggests the offensive line wasn't really the problem last year:
The two columns at right have two different ways of calculating sack rates. In the end, it doesn't really change the numbers much, although I think the big difference between Johnson's 07 and 08 numbers probably comes from kneeldowns. The biggest difference in overall sack rates just comes from Johnson playing more last year.
For the overall shape of the passing offense, we have this chart (click for full size):
I cut it down to the top players at each position because it removes a lot of clutter. The bottom line gives you the contributions of the chaff.
Basically, it looks to me like the Cowboys had two good receivers (2005), two very good receivers (2006), and then just Terrell Owens and a bunch of guys (2007 and 2008). I know the folks in Dallas are all excited about Miles Austin -- and he did post an impressive YPC of 21.4 last year -- but even with the addition of Roy Williams, I don't see how this isn't a much less talented receiving corps than they used to have. Patrick Crayton looks like a #2 Eagles receiver (pre-2008 edition), meaning he'd be a #3 anywhere else.
With that said, Dallas may be making it up somewhat in other areas. After years of watching how the Eagles use Brian Westbrook, it shouldn't be hard for Dallas to figure out how to get Felix Jones involved in the passing game, if the receivers don't step up.
There's nothing really interesting about Romo's stats last year. He was pretty much the same player, just not quite as good. Blame Jessica, TO, Garrett or a decrease in receiving talent (including the aging of TO), as you see fit.
The last question, and maybe the most important, is what we can expect from Roy Williams this year. I'm honestly not that impressed by the guy. He's got great physical attributes, but on the field he seems like a poor man's Braylon Edwards. Compare his last four seasons to TO's:
Now, granted, you can't use these stats to strip out quarterback effects. There's no question about who was in the better situation. But even in Williams' big-stat 2006 season, his FO numbers just aren't that good. A DYAR of 250 and DVOA of 8.7% puts you just outside the top 10 in terms of NFL receivers. That's down in Derrick Mason / Greg Jennings territory. Those guys are good receivers, but they're not HOFers. And while Williams' big body can provide matchup issues in the red zone, he's not a guy with the greatest hands. He'll make some amazing catches, but he'll also drop some he should come up with. You at least have a shot down there.
After looking at all these numbers, I'm pretty firmly convinced that if the Cowboys are going to get the offense revving again, it's not going to be the receiving corps carrying the load. It's going to be on that big offensive line and a very talented backfield instead. And if Dallas is controlling the line of scrimmage and Romo has time when they pick his spots, they can be a very dangerous offense indeed.