- Messages
- 79,281
- Reaction score
- 45,652
8:00 AM Fri, Oct 17, 2008 | Permalink | Yahoo! Buzz
Tim MacMahon E-mail News tips
Bill Barnwell, a contributing editor at footballoutsiders.com and one of the minds behind the must-read Pro Football Prospectus, will answer five Cowboys questions each week of the season using the FO data and methodologies. Here's the Week 7 edition:
1. The Cowboys are convinced that Roy Williams is a major upgrade at the No. 2 receiver spot. What do the FO statistical formulas say?
According to his performance in the Detroit offense, not really. Williams' DVOA is a mediocre -9.6%, 57th in the league; he's caught only 44% of the 39 passes thrown to him. Now, admittedly, Williams has had Jon Kitna and Dan Orlovsky for quarterbacks, so he should see some improvement.
That being said, it'd hard to see where Williams is going to steal passes from. The idea that Patrick Crayton wasn't a good enough #2 receiver is kinda silly; he was good enough last year, when the Cowboys had one of the best passing attacks in football, and his 17.8% DVOA is 22nd in the league so far this year. Williams moves Crayton into the third slot, which moves Miles Austin basically out of the picture altogether. Austin, as mentioned previously, has been fantastic when thrown the ball. It also means that Jason Witten will see fewer touches, and he's been the best tight end in the league for the past year and six weeks.
The Cowboys aren't going to entirely overhaul their offense and become a spread attack in the middle of the season, especially with Brad Johnson potentially at the helm. That's going to result in everyone taking fewer touches to accommodate each other's talents. I wonder who that won't go over well with.
2. Based on Brad Johnson's performance during his last stint with the Vikings, should Cowboys fans be pressing the panic button?
It depends on what Brad Johnson you're getting. It's easy to point to the 2005 Brad Johnson that threw four interceptions in 294 attempts and say that he'll just "manage the game" and check down and avoid mistakes. Of course, that 1.4% interception ratio (interceptions divided by attempts) was way out of line with his averages and bounced back up dramatically the year after, when he threw 15 interceptions in 439 attempts (3.4% interception ratio).
I think what you're going to see with Brad Johnson is a lot more drives ending or stalling because of sacks. Tony Romo's ability to throw as effectively on the run as he does is rare; the fact is that he has to do it so frequently attests to the fact that the Cowboys pass blocking isn't as good as it's perceived to be with Romo behind center.
I'm not saying that Johnson will be Drew Bledsoe-level statuesque. He just won't be Romo.
As for the idea of having Romo start, it seems like a bad one. If you look back to last year's game against the Eagles in Week 15, Romo hurt his finger and/or his hand and was an absolute mess. It was his worst game as a professional, and although he wasn't the only problem in that game, his performance directly prevented the Cowboys from winning.
3. What's with the Cowboys' sudden burst of injuries?
In this year's Pro Football Prospectus, we pointed out how infrequently the Cowboys had suffered from these sort of medium-term injuries to their starters, and how unlikely that was to continue. The Cowboys' medical staff has done amazing work over the past few seasons, but it seemed impossible that they could defy the law of averages for another year.
Obviously, last year, the Cowboys had Jason Ferguson and Terry Glenn go down for the entire regular season, but there wasn't the sort of medium-term injuries, the out for 4-6 week injuries that we're seeing this year with Newman, the other Roy Williams, Kosier, and now (theoretically) Romo.
4. Who is the bigger loss: Felix Jones or Adam "Pacman" Jones?
I think Pacman. He hasn't been otherworldly this year like he was in 2006 with Tennessee, and Jones has been fantastic whenever he touches the ball, but the Cowboys are already loaded with playmakers. Without Newman, they're certainly not loaded with cover corners. I get the feeling that people are going to get very angry at Mike Jenkins until Terence Newman comes back.
5. Per the FO playoff odds, the Cowboys are likely to be left out of the postseason. Can you explain why the odds of the Cowboys making the playoffs are only 34.8 percent?
It's a combination of things. The way that those playoffs odds are calculated is by using DAVE, which our statistic that takes into account both a team's current DVOA as well as its preseason projected DVOA, with the latter becoming a smaller part of the equation each week. The idea is to mitigate some of the fluke factor of early-season teams while using our preseason projections, which have been more accurate than anyone else's since we started publishing them. We then use DAVE to simulate the season 10,000 times, and then list each team's average wins and the percentage of the time they made the playoffs. That leaves the Cowboys at 34.8%, likely owing to that low preseason projection and a tough schedule.
Of course, I still think that our preseason projection for the Cowboys underrates them some. Their DAVE is way below the rest of the division's, but we're projecting the difference between first and last place in the East to be only 1.3 wins -- if the Cowboys are really better than that DAVE number, and I think they are, their chances of making the playoffs could very easily rise dramatically.
What we can use on the other hand is a dumb model that only takes into account two things -- wins and games played. We know the Cowboys are 4-2, so what if we only use that as the data point and look at the past?
We know that, on average, teams that are 4-2 after six games win an average of 9.5 games. Teams with four wins in their first six games have a 62.5% chance of making the playoffs.
Using this model, we can denote the relative importance of a game contextually; if the Cowboys win this week, for example, they'd be 5-2 through seven games; teams with that record in the past have averaged 10.2 wins per season, and make the playoffs 76.7% of the time. That would mean that a win against St. Louis would boost the Cowboys' playoff chances by 13.2%.
On the flip side, a loss to the lowly Rams would hurt the Cowboys' chances dramatically. Teams that are 4-3 through seven games average only 8.6 wins and make the playoffs less than half the time, at 49%. That would cause the Cowboys' expectations of making the playoffs to fall by 13.5%.
Obviously, take these constructions with a grain of salt. There's still a whole lot of time between now and January.
Tim MacMahon E-mail News tips
Bill Barnwell, a contributing editor at footballoutsiders.com and one of the minds behind the must-read Pro Football Prospectus, will answer five Cowboys questions each week of the season using the FO data and methodologies. Here's the Week 7 edition:
1. The Cowboys are convinced that Roy Williams is a major upgrade at the No. 2 receiver spot. What do the FO statistical formulas say?
According to his performance in the Detroit offense, not really. Williams' DVOA is a mediocre -9.6%, 57th in the league; he's caught only 44% of the 39 passes thrown to him. Now, admittedly, Williams has had Jon Kitna and Dan Orlovsky for quarterbacks, so he should see some improvement.
That being said, it'd hard to see where Williams is going to steal passes from. The idea that Patrick Crayton wasn't a good enough #2 receiver is kinda silly; he was good enough last year, when the Cowboys had one of the best passing attacks in football, and his 17.8% DVOA is 22nd in the league so far this year. Williams moves Crayton into the third slot, which moves Miles Austin basically out of the picture altogether. Austin, as mentioned previously, has been fantastic when thrown the ball. It also means that Jason Witten will see fewer touches, and he's been the best tight end in the league for the past year and six weeks.
The Cowboys aren't going to entirely overhaul their offense and become a spread attack in the middle of the season, especially with Brad Johnson potentially at the helm. That's going to result in everyone taking fewer touches to accommodate each other's talents. I wonder who that won't go over well with.
2. Based on Brad Johnson's performance during his last stint with the Vikings, should Cowboys fans be pressing the panic button?
It depends on what Brad Johnson you're getting. It's easy to point to the 2005 Brad Johnson that threw four interceptions in 294 attempts and say that he'll just "manage the game" and check down and avoid mistakes. Of course, that 1.4% interception ratio (interceptions divided by attempts) was way out of line with his averages and bounced back up dramatically the year after, when he threw 15 interceptions in 439 attempts (3.4% interception ratio).
I think what you're going to see with Brad Johnson is a lot more drives ending or stalling because of sacks. Tony Romo's ability to throw as effectively on the run as he does is rare; the fact is that he has to do it so frequently attests to the fact that the Cowboys pass blocking isn't as good as it's perceived to be with Romo behind center.
I'm not saying that Johnson will be Drew Bledsoe-level statuesque. He just won't be Romo.
As for the idea of having Romo start, it seems like a bad one. If you look back to last year's game against the Eagles in Week 15, Romo hurt his finger and/or his hand and was an absolute mess. It was his worst game as a professional, and although he wasn't the only problem in that game, his performance directly prevented the Cowboys from winning.
3. What's with the Cowboys' sudden burst of injuries?
In this year's Pro Football Prospectus, we pointed out how infrequently the Cowboys had suffered from these sort of medium-term injuries to their starters, and how unlikely that was to continue. The Cowboys' medical staff has done amazing work over the past few seasons, but it seemed impossible that they could defy the law of averages for another year.
Obviously, last year, the Cowboys had Jason Ferguson and Terry Glenn go down for the entire regular season, but there wasn't the sort of medium-term injuries, the out for 4-6 week injuries that we're seeing this year with Newman, the other Roy Williams, Kosier, and now (theoretically) Romo.
4. Who is the bigger loss: Felix Jones or Adam "Pacman" Jones?
I think Pacman. He hasn't been otherworldly this year like he was in 2006 with Tennessee, and Jones has been fantastic whenever he touches the ball, but the Cowboys are already loaded with playmakers. Without Newman, they're certainly not loaded with cover corners. I get the feeling that people are going to get very angry at Mike Jenkins until Terence Newman comes back.
5. Per the FO playoff odds, the Cowboys are likely to be left out of the postseason. Can you explain why the odds of the Cowboys making the playoffs are only 34.8 percent?
It's a combination of things. The way that those playoffs odds are calculated is by using DAVE, which our statistic that takes into account both a team's current DVOA as well as its preseason projected DVOA, with the latter becoming a smaller part of the equation each week. The idea is to mitigate some of the fluke factor of early-season teams while using our preseason projections, which have been more accurate than anyone else's since we started publishing them. We then use DAVE to simulate the season 10,000 times, and then list each team's average wins and the percentage of the time they made the playoffs. That leaves the Cowboys at 34.8%, likely owing to that low preseason projection and a tough schedule.
Of course, I still think that our preseason projection for the Cowboys underrates them some. Their DAVE is way below the rest of the division's, but we're projecting the difference between first and last place in the East to be only 1.3 wins -- if the Cowboys are really better than that DAVE number, and I think they are, their chances of making the playoffs could very easily rise dramatically.
What we can use on the other hand is a dumb model that only takes into account two things -- wins and games played. We know the Cowboys are 4-2, so what if we only use that as the data point and look at the past?
We know that, on average, teams that are 4-2 after six games win an average of 9.5 games. Teams with four wins in their first six games have a 62.5% chance of making the playoffs.
Using this model, we can denote the relative importance of a game contextually; if the Cowboys win this week, for example, they'd be 5-2 through seven games; teams with that record in the past have averaged 10.2 wins per season, and make the playoffs 76.7% of the time. That would mean that a win against St. Louis would boost the Cowboys' playoff chances by 13.2%.
On the flip side, a loss to the lowly Rams would hurt the Cowboys' chances dramatically. Teams that are 4-3 through seven games average only 8.6 wins and make the playoffs less than half the time, at 49%. That would cause the Cowboys' expectations of making the playoffs to fall by 13.5%.
Obviously, take these constructions with a grain of salt. There's still a whole lot of time between now and January.
