Writers of Pro Football Prospectus 2008

Most Recent FO Features

RiversPhi04.jpg

» Film Room: Philip Rivers

The "blueprint" the Chargers used to beat the Seahawks? It starts with a Pro Bowl quarterback playing at the absolute top of his game.

22 Feb 2010

Four Downs: AFC South

by Vince Verhei and Tom Gower

(Editor's note: This was originally published with the first draft of the Indianapolis "Who Could Leave?" section. That section has since been updated.)

Houston: Can Steve Slaton bounce back?

Steve Slaton's rookie season was the stuff dreams are made of. His 1,282 rushing yards set a Texans franchise record and led all first-year players in 2008. Those dreams turned to nightmares in 2009, though, as Slaton rushed for just 437 yards and 3.3 yards per carry. He also fumbled five times in just 131 carries. (Adrian Peterson led all running backs with six fumbles, but he also carried the ball 314 times). Slaton was finally placed on Injured Reserve after Week 12 with a neck injury.

Slaton was even worse by Football Outsiders' advanced metrics, finishing last in the league in all three statistics we use to evaluate running backs: DVOA (Defense-adjusted Value Over Average, which measures value on a per-carry basis), DYAR (Defense-adjusted Yards Above Replacement, which measures total value), and Success Rate (which measures how often a runner gains meaningful yardage). Although his speed and open-field ability made him a very effective receiver (he finished third among running backs in receiving DVOA, and second in receiving DYAR), Slaton was pretty clearly the most impotent ground threat in football in 2009.

And the problem wasn't the Texans' offensive line, it was definitely Slaton; his teammate Ryan Moats ranked second in the NFL in Success Rate, and Arian Foster would have finished second in the league in DVOA if he had carried the ball often enough to qualify.

Then came word that Slaton underwent neck surgery in January. Slaton called the procedure a "cervical fusion," while Texans coach Gary Kubiak used the term "discectomy." Regardless, the procedure performed by Dallas surgeon Drew Dossett was to relieve pressure on the nerve root of the spine. In a chat on the Texans' Web site, Slaton suggested that he had been hurt all season: "You always have aches and pains as a player," he wrote, "but this got progressively worse." He specifically blamed the injury for his fumbles, saying his right side had been weaker than his left.

Though Slaton is claiming he'll be stronger than ever by the draft, there's no guarantee he'll ever return to his rookie form. Our own injury expert Will Carroll could think of only one player who has returned a year after a similar procedure: Brad Johnson, a quarterback who played 11 more seasons and won a Super Bowl after suffering a herniated cervical disk in 1997. As a running back, however, Slaton will be subjected to more of a pounding than Johnson was. Slaton is wading into largely uncharted waters.

Who could leave?

The biggest name among Houston's unrestricted free agents is cornerback Dunta Robinson. Robinson was also a free agent last season, but Houston retained him via the franchise tag. Houston Chronicle writer John McClain has written that Robinson will leave for sure if he's not franchised again this offseason. The Texans got plenty of quantity from Robinson last season (he played every game for the first time since 2006), but not much quality, as his level of play slipped badly.

Kevin Walter, the Wes Welker to Andre Johnson's Randy Moss, is also up for grabs; he has been in the top 10 in receiving DVOA each of the past two seasons.

The Texans' punt team ranked fifth in our ratings last year, and while that's partly because their coverage teams allowed only 4.3 yards per return, punter Matt Turk also deserves a lot of credit. He and Washington's Hunter Smith were the only two players with more than 10 punts last season to force as many fair catches as they allowed returns.

The Texans were bailed out by the collapse of the Collective Bargaining Agreement. Under the old plan, three more starters -- tight end Owen Daniels, safety Bernard Pollard, and star linebacker DeMeco Ryans -- would have also been unrestricted free agents. Instead, while they are free agents, they are restricted and not likely to be going anywhere.

Who could they sign?

The Texans should look to the Super Bowl champions to fill the biggest question mark on their roster. The Saints will likely retain restricted free agent running back Pierre Thomas, but Mike Bell and his 54 percent Success Rate (sixth in the league) could be available. Whether Robinson leaves or not, the Texans will be looking for cornerbacks. Some of the best names available include Leigh Bodden, Tramon Williams, and likely the best of the bunch, Carlos Rogers.

Indianapolis: Can the Colts fix their running game?

It's hard to find flaws on a team that went 16-1 in meaningful games last year, but the 2009 Colts had one obvious weakness: They couldn't run the ball for squat, finishing with only 1,294 rushing yards and 69 first downs on the ground, both the lowest figures in the league. That's largely because they didn't rush very often -- they ended up with 366 carries, one more than Arizona and less than anyone else -- but the advanced metrics at Football Outsiders show this was still the biggest hole on the team. The Colts offense finished 22nd in rushing DVOA (Defense-adjusted Value Over Average, which measures value on a per-carry basis).

The problem was not with Joseph Addai, who was 17th out of 50 individual running backs in DVOA. Instead, it was with backup runners Donald Brown and Mike Hart, who would have ranked 44th and 46th in DVOA if they had garnered enough carries to qualify for the leader board.

The offensive line deserves some of the blame too. The Colts ranked 25th in Adjusted Line Yards, Football Outsiders' metric that attempts to separate the performance of an offensive line from its running backs. While center Jeff Saturday is still excellent, his linemates are subpar. The Colts were eighth in ALY in runs up the middle, but no higher than 19th in any other direction. Injuries had little to do with the line's performance -- the starters combined to miss 16 games, which is about average for most teams.

It's doubtful that an impact offensive lineman will be available when the Colts pick 31st in the first round of April's draft, but there should be some good talent at running back, possibly including Jahvid Best out of California. Or the Colts could cross their fingers and hope that Donald Brown takes a giant step forward in his sophomore season. Either way, there about 30 other teams who wish their biggest offensive problem entering 2010 was "backup running back."

Who could leave?

The biggest question facing the Colts in free agency is the future of Gary Brackett. The five-year starter is an unrestricted free agent, and while both player and team say they want Brackett to stay, there's no guarantee a deal will get done. As Mike Chappell of the Indianapolis Star points out, the Colts have never invested a big contract to retain a free agent-to-be linebacker (see June, Cato).

The Colts have a number of players up for restricted free agency, including offensive tackle Charlie Johnson, defensive tackle Antonio Johnson, and linebackers Tyjuan Hagler and Freddy Keiaho. If Brackett leaves, Keiaho then returns to the starting lineup. Similarly, if Charlie Johnson signs elsewhere, former starter Tony Ugoh will find himself on the first string once more.

Kicker Matt Stover is also an unrestricted free agent. With Adam Vinatieri under contract and returning to health, the Colts will wish Stover the best in his future endeavors.

Who could they sign?

Every year, this is the easiest offseason write-up we have to do. The Colts simply do not throw money at free agents. Of the 44 Colts who played in the Super Bowl against the Saints, only four had ever played on another NFL team, including just one starter (Antonio Johnson). The Colts stockpile talent in the draft or pick up undrafted rookies on the cheap, and save their money to retain players like Peyton Manning, Reggie Wayne, and Dwight Freeney.

Jacksonville: Are the Jaguars looking for a new quarterback?

The list of primary starters at quarterback for the Jacksonville Jaguars from their inception reads Mark Brunell, Byron Leftwich, and David Garrard. Head coach Jack Del Rio declared Garrard a good but not elite quarterback and one who needs a strong team around him to reach the Super Bowl.

It was a bold statement from the coach who committed to Garrard over high draft pick Byron Leftwich on the eve of the 2007 season and signed Garrard to a six-year, $60 million extension after that year. Garrard played like that in 2007, finishing third in DVOA and seventh in DYAR, but slipped to 15th and 14th in 2008 and 23rd and 19th in 2009. Offensive line struggles contributed to those lesser numbers, as did lack of production from wide receivers, but the Jaguars were one of the league’s most inconsistent offenses thanks in large part to Garrard’s inconsistent play.

Garrard turned 32 on Valentine’s Day, so he is unlikely to improve his level of performance much unless the talent around him improves. Improvement is reasonable to expect from the offensive line, as 2009 draftees Ebon Britton and Eugene Monroe gain experience, but despite Mike Sims-Walker’s emergence, wide receiver remains a weak link. That makes a Super Bowl run unlikely in the Garrard era.

With the 10th or 11th pick in the draft (depending on a coin flip with the Bears) and Garrard in place, the Jaguars should be well-positioned to take Sam Bradford or Jimmy Clausen and give him a year to adapt to the NFL rather than thrusting him into an immediate starting role. The fans in Jacksonville might prefer hometown hero Tim Tebow, but the fans in Los Angeles will appreciate the front office’s patience and foresight.

Who could leave?

Jacksonville kicked things off early by releasing wideout Torry Holt, tackle Tra Thomas, and defensive tackle Rob Meier. Beyond them, the Jaguars have a particularly short list of impending free agents. Only Kynan Forney, Reggie Hayward, and Ernest Wilford will be unrestricted. Given the disappointments that 2008 draftees Derrick Harvey and Quentin Groves have been, the Jaguars will look to retain Hayward. Beyond them, safety Reggie Nelson is a candidate to be released after being a disappointment the last couple years.

Who could they sign?

The Jaguars should be modest players in free agency. One way to make a splash without spending too much money would be to sign Darren Sharper to provide needed veteran support to a pass defense that ranked 32nd in DVOA. They could also address the wide receiver position, as Holt’s signing didn’t work out as well as expected. Kevin Walter might be an attractive complement to Mike Sims-Walker. An experienced backup for Maurice Jones-Drew is another possibility.

Tennessee: Can the Titans rebuild their defense?

The Titans made the playoffs in 2007 and 2008 behind defenses ranked first and fifth in DVOA, respectively. But with the departure of Albert Haynesworth and the aging of other defensive starters -- including end Kyle Vanden Bosch, outside linebackers Keith Bulluck and David Thornton, and cornerback Nick Harper -- the defense slipped to 27th last year. The fall was particularly acute in pass defense, as five of the first six quarterbacks the Titans faced passed for more than 300 yards. The poor pass defense helped put the team in an 0-6 hole they couldn't dig out of.

The Titans’ depth behind Bulluck, Thornton, and Harper was tested last year when each went down with injuries. Although rookie linebacker Stanford Keglar proved adequate, neither rookie corner, Ryan Mouton or Jason McCourty, looked prepared to assume the starting job created by Harper’s free agency.

An improved pass rush would help the secondary, as the Titans fell from ninth in the league in Adjusted Sack Rate (sacks per pass attempt adjusted for opponent, down, and distance) in 2008 to 25th in 2009. Veteran ends Vanden Bosch and Jevon Kearse have each lost a step and will likely be allowed to leave. William Hayes largely stepped into Kearse’s shoes in the second half of 2009, but the Titans may still look to address the position in free agency. A healthy season from defensive tackle Jason Jones would also help.

Who could leave?

Vanden Bosch, Kearse, Bulluck, and Harper are all unrestricted free agents, and each is unlikely to return. Thornton’s sizable base salary and recent injury history could also lead to his release if the team feels a suitable replacement is on the roster. On the offensive side of the ball, pending free agent Eugene Amano should be re-signed, which means Pro Bowl center Kevin Mawae will be allowed to walk. Quarterback Kerry Collins, a backup with a starter’s salary, may also find himself cast loose if the Titans add another passer. The Titans may be willing to part with running back LenDale White, who has expressed an interest in more than the backup role he’d probably play again in 2010.

Who could they sign?

Defensive end, outside linebacker, and cornerback are the positions the Titans seem most likely to address. Julius Peppers and Karlos Dansby will be too expensive for the Titans’ mid-market blood, but a mid-level player like Pisa Tinoisamoa or a new contract for Roderick Hood might be an option.

(Portions of this article originally appeared on ESPN.com Insider.)

Posted by: Tom Gower on 22 Feb 2010

55 comments, Last at 01 Mar 2010, 10:20am by DeltaWhiskey

Comments

1
by Thunderbolt of ... :: Mon, 02/22/2010 - 12:29pm

And with one sentence, FO loses 95% of its readership in northern Florida.

2
by Dean :: Mon, 02/22/2010 - 12:34pm

That would imply that they have 20 readers there. Which would imply that the Jags have 20 fans.

By all accounts, the fans they do have are great, but sadly it seems there's just not enough of them.

3
by Tom Gower :: Mon, 02/22/2010 - 1:10pm

And, of course, the Titans have officially re-signed Eugene Amano since the language about re-signing him was submitted.

4
by Jason :: Mon, 02/22/2010 - 1:21pm

Re: Colts team composition—Antonio Johnson officially never played for another team. He was on the Titans' practice squad until the Colts signed him in the middle of the 2008 season.

Who has two thumbs and is still bitter about losing him? This guy.

If I hit a grand slam on this hole-in-one the dominoes will fall like a house of cards. Checkmate.

5
by nat :: Mon, 02/22/2010 - 1:50pm

...the 2009 Colts had one obvious weakness: They couldn't run the ball for squat.

Special teams?

The running game stats can be skewed by how and when the Colts run, and by the last game-and-a-half of the regular season. Special teams stats, not so much.

And did you see the Super Bowl? I heard it was on TV this year.

7
by Paul R :: Mon, 02/22/2010 - 2:15pm

I watched the Super Bowl.
The Colts special teams must be good if the coach is calling for a 52-yard field goal.
He obviously doesn't have any confidence in the team's passing game. Maybe they should get some help there.

13
by Bobman :: Mon, 02/22/2010 - 5:10pm

Overall, the ST improved from 2008 to 2009, but most of that was due to MacAfee kicking off and Aaron Francisco tackling--when he missed a handful of late games, things slid a bit. As Tanier pointed out a few weeks ago, the blocking on returns seems poor, but the coverage is pretty good.

I don't THINK they need a major overhaul, but would not mind a 2nd round pick spent on a top-flight return man who can also play dime CB (or 6th WR) if need be....

6
by jedmarshall :: Mon, 02/22/2010 - 2:12pm

Brown was ok before he got injured and think he needs another chance next year to prove himself. I certainly think there would be better ways to use the 31st pick than another back. The line is a bigger problem, especially run blocking. Addai had some of the hardest 3-yard runs in the league this year and was lucky to get back to the line of scrimmage a lot of times when he was hit almost upon taking the handoff.

They don't appear to be as bad as they are become Peyton and Saturday are so good, they can mask the weaknesses to a degree. They are especially bad at guard and only adequate at tackle.

Brackett will be re-signed and I'd expect Bethea to be as well.

8
by sam :: Mon, 02/22/2010 - 2:27pm

1. Reggie Nelson was not a disappointment the last few years. He was absolutely terrible.

2. They could also address the wide receiver position, as Holt’s signing didn’t work out as well as expected. Kevin Walter might be an attractive complement to Mike Sims-Walker.
The team drafted three wide receivers last year and (I think) still likes what Troy Williamson might have to offer in addition to Mike Sims-Walker. I'd almost guarantee they won't sign Walter or any other name-recognition Unrestricted Free Agent WR. Further, Holt's signing worked out pretty much exactly like they'd hoped: he was only ever an insurance policy in case Mike Sims-Walker couldn't get healthy and if they didn't find anybody in the draft. They never expected him to be more than a 3rd dow/hands-guy at his age and coming off the knee injury. The fact they released Holt I think tells you that they don't need to sign somebody else.

3. An experienced backup for Maurice Jones-Drew is another possibility.
That strikes me as extremely unlikely with Greg Jones and Rashad Jennings on the roster. I just don't see them spending money on a backup free agent RB.

--
sam! or the original sam from the old FO

32
by iapetus (not verified) :: Tue, 02/23/2010 - 10:57am

There's only one real possibility, and that's BAP. Drafting (badly) for need hurt the Jaguars over the last few years. That isn't Smith's approach, though, and it isn't what the team needs - it's not like there's just a couple of holes to fill to make them a regular Superbowl contender...

9
by Vincent Verhei :: Mon, 02/22/2010 - 3:01pm

The "Who Could Leave" section for the Colts has been updated. An early-draft version of that section was accidentally published.

10
by Jimmy :: Mon, 02/22/2010 - 3:17pm

Can any Titans fans give any insight as to how much Bulluck has left?

11
by Independent George :: Mon, 02/22/2010 - 4:30pm

The problem was not with Joseph Addai, who was 17th out of 50 individual running backs in DVOA. Instead, it was with backup runners Donald Brown and Mike Hart, who would have ranked 44th and 46th in DVOA if they had garnered enough carries to qualify for the leader board.

Goddamnit, Donald.

12
by nat :: Mon, 02/22/2010 - 4:45pm

Instead, it was with backup runners Donald Brown and Mike Hart, who would have ranked 44th and 46th in DVOA

According to DYAR, Peyton Manning did as much damage to the Colts running game in 19 carries as these two guys did in 107 carries. (Although comparing is inexact, because QB running DYARs are different from backs' DYARs) Not that I would replace Manning to improve the running game. (Vick, maybe? I thought not.)

14
by Bobman :: Mon, 02/22/2010 - 5:11pm

Kneel to win, baby. More proof that the run game is deader than... dead, man.

17
by tuluse :: Mon, 02/22/2010 - 6:14pm

It's dead for teams that have Peyton Manning.

For the rest of us, it's still nice to have.

38
by Bobman :: Tue, 02/23/2010 - 3:47pm

Tuluse,

True. #1 I was kidding, of course, and #2 paraphrasing Pulp Fiction. (Now comparing the run game to cocaine... that's a little strange, I admit.)

30
by nat :: Tue, 02/23/2010 - 9:28am

I knew that DYAR and DVOA doesn't include kneeldowns, but I didn't realize that FO still included kneels in the play count.

I can't find a single called run play for Manning in the regualr season. He had two scrambles that would count as runs: 3 yards on second down and six to go, and 2 yards on third and three. Failing to get a first down in that last case probably counts a lot, since QBs run there only if they see they can get the first down, and any tackle for a loss counts as a pass play in DVOA.

Those two runs resulted in -21 DYAR, but I can't fault the running game at all for those.

35
by Eddo :: Tue, 02/23/2010 - 2:20pm

I recall (though this could have been in 2008) Quick Reads discussing a failed QB sneak of Manning's counting for more than -20 DYAR.

QB rushing DYAR seems relatively useless to me, in general, unless a QB has at least 50 rushing attempts. Otherwise, there are too many extreme results; a failed QB sneak counts very negatively, while a non-running QB might be able to scramble for a first down on third-and-long due to the defense not expecting it and having everyone in coverage.

45
by nat :: Wed, 02/24/2010 - 12:50pm

Week 11 Quick Reads had Tom Brady scored at -22 DYAR rushing for going 1-for-3 on short yardage rushes and a scramble that did not get a first down. That may be what you are thinking of.

Brady...Manning - it's a common mistake.

46
by Eddo :: Wed, 02/24/2010 - 4:09pm

Could be. Seems likely.

The important thing is I didn't imagine it.

15
by Bobman :: Mon, 02/22/2010 - 5:14pm

Oh, and am I the only stunned by the news that Garrard is 32? What was he, born in his 20s? My impression of him was that he spent about 2-3 years behind Leftwich, and two years ago took the helm. Clearly this is wrong, as it suggests he's maybe 26-27 or so. 32? really? Brady's age? A year younger than P Manning and 5 years older than Eli? I have to pay closer attention....

16
by Shake (not verified) :: Mon, 02/22/2010 - 6:00pm

Entered the league at 24, 3 1/2 years behind Leftwich, took over mid 2006.

Tony Romo is the QB whose age always surprises me.

21
by HostileGospel :: Mon, 02/22/2010 - 10:04pm

Right there with you, Bobman. The funny thing is that I know I've read stories about him before, and I remember his age being mentioned, but I must not have processed it properly or something.

--
There's a place I want to be. It's the NovaCare Center. That's in Philadelphia. One NovaCare Way, where the Eagles practice and then they eat cafeteria food and they watch film and we eat and we have fun.

-Donovan McNabb

18
by commissionerleaf :: Mon, 02/22/2010 - 6:29pm

Colts:

The Colts have to tread lightly in retaining Brackett and others because they have the Manning renegotiation coming up. I would expect them to let Brackett walk; he's a good MLB, and not an easy player for them to replace, but he's not special or essential to what the Colts do the way Mathis, Freeney, and Manning are. Bethea comes cheaper if they want to keep a veteran presence, and there is probably a fast, undersized, smart linebacker that Bill Polian can find with the 31 pick.

In terms of needs, there are many. The defensive secondary has been rocked by injuries; Bob Sanders clearly can't be relied upon and Bethea may or may not return. The cornerbacks were essentially in rotation because apart from Hayden, they were all new to the NFL. And Hayden has never been good (although everyone was better than Tim Jennings, who needs a pink slip). Personally, I'd love to see Session and Bullitt gone, because they're both bad tacklers (at FS and LB!), but the team can't replace everyone at once. And the team needs line help too on offense, although the skill positions look set.

I'd look to add a Sanders replacement or, if someone looked interesting, an eventual Freeney/Mathis replacement in the draft after replacing Brackett, try to get Bethea resigned, and sit tight and hope the team can hold up another couple of years before the Manning/Freeney/Mathis decline ends the dynasty.

Texans:

Lots of needs, but some good pieces. Letting Robinson and Walter go is probably right, and adding young talent in the defensive secondary wouldn't be a mistake. It's probably too much money, but if the Rams were serious about trading down, Suh would look great next to Williams on the defensive line.

Titans:

The defense fell apart without Haynesworth. Needs a bottom-up rebuild now that every good player from the mid-decade defensive unit is gone or declining. Draft defenders, and sign any good D lineman available.

Jags:

This team's problems go back to the ill fated decision to blow an entire draft class on failed pass rushing talents. The lines are probably okay, but the rest of the team has declined. At this point, it's proabbly best to go into rebuilding mode, draft a young QB, and see where he takes you.

19
by Tom Gower :: Mon, 02/22/2010 - 7:06pm

Jaguars first round picks under Shack Harris:
2003: Byron Leftwich, 2004: Reggie Williams, 2005: Matt Jones, 2006: Marcedes Lewis, 2007: Reggie Nelson, 2008: Derrick Harvey
After getting axed by the Jaguars, Shack Harris was hired by the Detroit Lions. I wish I was making that up.

20
by The Human Spider :: Mon, 02/22/2010 - 7:54pm

Sounds like the Detroit Lions to me.

28
by iapetus (not verified) :: Tue, 02/23/2010 - 7:32am

And what exactly do you think the Jags were doing last year? You know, when they parted ways with guys like Fred Taylor and started the rebuilding with young tackles in back-to-back rounds?

You draft a young QB when a young QB's available in the draft. They didn't get their guy last year - if there's one available this year where they pick, then they probably will.

Make no mistake, this team is already rebuilding.

39
by Bobman :: Tue, 02/23/2010 - 3:51pm

Commish: Session gone! He was their top tackler, not that that proves much--he still could have missed a million chances. But my impression is that he's greatly improved and just entering his prime.

I'd keep Brackett until his wheels fall off, assuming the money is reasonable. He SEEMS like he's in the Rob Morris mold as a guy who will take getting replaced some day gracefully and lead from the bench for a year or two, plus fill in with veteran savvy when somebody goes down.

48
by commissionerleaf :: Wed, 02/24/2010 - 6:59pm

On Sessions: I simply do not understand the idea of a starting linebacker who leads the league in missed tackles (I think Sessions and Bullitt were 1-2, actually... anyway both of them were in the top 5). It's so damaging to have drives extended even though you schemed correctly. I recognize that Indy does well with undersized guys, and that missed tackles can be the result of that, but Sessions is really problematic. I suspect that the tackle numbers may have something to do with teams running at him on purpose, which perhaps isn't as big a problem at weakside LB in 2009 as it was as the SSLB in '08. Still, he's not that great in coverage and he isn't a pass rusher (not that he needs to be, I recognize, in the Colts system - though that's changing somewhat now). Session isn't a bad player, but he's a guy the Colts need to look to replace, unless they have some other strategy for improving against the run (Session improved in coverage a lot since his first year).

On Brackett: I agree Brackett is great, and very much in the Colts mold. But if he wants Brian Urlacher money, they need to let him go; I don't want to see them letting Bethea or, God Forbid, a lineman, go (or go by) because they're paying an MLB. The Colts defense isn't as simple as it was in the Dungy era, but it's still not Baltimore; having a veteran MLB presence isn't essential.

23
by Jeff Fogle :: Tue, 02/23/2010 - 1:37am

The Indianapolis Colts don't have a broken running game. They use a short passing game as a virtual running game as they move the chains and drive the field. It's a choice they're making. They used this approach to go 16-1 in games that mattered, while leading the league in points scored on drives of 60 yards or more with a little over 18 per game (about a TD better than the league midpoint), and leading the league in third down conversions at 49% (league midpoint was about 40%).

This choice allows them to stay way ahead of the risk/reward curve. Quick hitters that move the chains don't just add up to volume in the good stuff, they help cut down on the bad stuff.

Indianapolis:
#1 in fewest offensive fumbles with 11 (league midpoint 24)
#1 in fewest lost fumbles with 5 (league midpoint 10)
#2 in fewest penalties with 74 (league midpoint 95)
#2 in fewest penalty yards with 546 (league midpoint in the 770-790 range)

This is a big deal. It helps define why they're so successful...why they earned the #1 seed in the AFC...then won both of their AFC playoff games by double digits. They didn't win the Super Bowl, but they did win 3rd downs 46% to 33%, and scoring on long drives 14-3 (and, for that matter, rushing yardage 99-51) over the Saints. Their defense didn't get much second half pressure on Brees...allowing him to have a very efficient game. That plus the pix six spelled doom.

The point is, the rushing game is not a "weakness" that needs to be "fixed." They've found a better approach (helped obviously by a quick-thinking QB in a smart schematic)that has made things more efficient. The fact that the "efficiency" that DVOA claims to measure doesn't catch this is a strike against DVOA. Indy's offense led the league in stuff that mattered, but didn't tack on superfluous noise. Not as good as their record? I'll buy that. They weren't passing out of desperation because they couldn't run though...they were using quick hitters as a virtual running game because it was more effective.

Back in 1980, Mike Schmidt of the Philadelphia Phillies won the NL MVP award with a batting average of just .286. He led the league in HR's with 48, and RBI with 121. He led the league in slugging percentage with .624, and OPS with 1.004. He was also a gold glove third baseman. Would a section header called "Can Mike Schmidt fix his batting average" been appropriate? Would something calling his batting average a hole reflect "advanced metrics?" Schmidt made a choice to emphasize power, while accepting walks when given (obp was .380).

Indy made a choice that put them ahead in the risk/reward curve in a way that moved the chains, put points on the board, and made them particularly dangerous in late game situations (with the exception of the SB of course, lol). It didn't ultimately win them the Lombardi Trophy. It sure gave them a great shot at it. Offensively, they had a Mike Schmidt season.

24
by DeltaWhiskey :: Tue, 02/23/2010 - 4:02am

"The point is, the rushing game is not a "weakness" that needs to be "fixed." They've found a better approach (helped obviously by a quick-thinking QB in a smart schematic)that has made things more efficient. The fact that the "efficiency" that DVOA claims to measure doesn't catch this is a strike against DVOA."

Your first point appears to be accurate. Your second, regarding a strike against DVOA, not so much. The problem is FO's interpretation of their data. DVOA says IND does not run particularly well, therefore, according to FO it is a weakness. You propose the alternative, that in fact IND has found a better replacement for the run game.

25
by Jerry :: Tue, 02/23/2010 - 4:16am

It may be that properly measuring Peyton Manning is as hard as imitating him (nobody in the NFL is copying the Colt offense despite its success). I'm sure that a tweak that would more accurately represent the Colts without negatively affecting other teams' DVOA would be implemented forthwith, but it's not worth being less accurate for most teams to get one more right.

Schmidt hit .316 in 1981 and won the MVP. His HR and RBI were proportionally as good or better than the year before ('81 was strike-shortened), and his OPS was better as well (1080). So he did, arguably, "fix his batting average", although it fell back down after that.

29
by DeltaWhiskey :: Tue, 02/23/2010 - 7:51am

This is not an issue of measuring P. Manning, but an issue of understanding him in relationship to the rest of the team and by extension the rest of the data. IND's run DVOA reflects an area of weakness statistically speaking, but in reality is not that important of a finding, except as it relates to understanding Manning better.

Warning slight tangent:

From a statistical standpoint we are talking about the difference in a finding that is statistically significant (which to be honest, we don't know if IND's poor run DVOA is significantly below the mean, but let's assume) and the importance of statistical significant findings. Frequently on this site, the cry is raised about sample size, and while often correct, the better approach is to perform the appropriate probability test for said statistic. This then allows one to make a confidence statement about said statistical finding. Moreover, the reality is, that if the sample size is increased sufficiently, there is often a good chance that a stastically significant finding can be found; however, this may be a truly meaningless finding if the magnitude and meaning of effect are ignored.

37
by Jeff Fogle :: Tue, 02/23/2010 - 3:04pm

Appreciate the responses. Regarding Schmidt, tough to say he "fixed" his batting average because, after the outlier year in BA, he spent the next 7 years below .295 again. '81 was the only season in his life where he topped .295...so I'm skeptical that he fixed batting average then immediately forgot what he fixed. His career batting average was .267, yet he was top ten in the MVP voting 9 times and won it 3 times. To me the general point holds. We could use '86 instead to make the point if you wish...where he again won all the same categories on the way to an MVP. I wanted to pick his best year, so I went with '80. The point is, batting average wasn't a weakness, it just wasn't a point of emphasis in his successful attack.

Regarding whether it's a strike against DVOA, I'd argue that it is. I think a very clear statistical case can be made that Indy had the most efficient offense in the NFL. They combined big volume with minimal mistakes in a way that constantly topped opponents.

Note, I went back and checked the numbers. Have to fix an error. Indy wasn't #1 in the league in points scored on long drives in the regular season. They were #2 behind New Orleans. It's a virtual dead heat if you count the playoffs though (51-38 for Indy over NO in 3 postseason games)

POINTS PER GAME ON DRIVES OF 60 YARDS PLUS (regular season)
1..New Orleans 19.4
2..Indy 18.3
3...Minnesota 17.5
4...New England 17.2
5...Green Bay 15.9
League midpoint: around 12
Crappiest teams: around 6
(These are done by hand with a spreadsheet, so it's not an official league stat, lol...just something I've kept by hand ever scoring drive lengths were listed in expanded boxscores...first done by USA Today in the pre-internet era)

I listed the negatives they avoided above (fumbles and penalties). I should have included sacks as well. Indy was #1 at avoiding sacks with only 13. The league midpoint was 34.

DVOA has:
Indy #6 on offense
Indy #9 in weighted offense
Indy #6 in passing offense
Indy #22 in rushing offense

I don't think that captures the season we just watched from the perspective of the Indy offense. It is the definition of efficiency at the team level, because of choices they've made about how to emphasize moving the chains and driving the field with a low risk/high reward approach.

Agree with you DW about the sample size stuff (and with most everything you've ever posted, lol). Only thing I'd add to the topic is the danger that evolutionary ticks or key changes present. There's always a short sample size just after an evolutionary tick happens...yet it does foreshadow the immediate future better than the prior sample size.

Quick illustration off the top of my head would be something like this. Wes Welker runs a 40 yard dash every weekday for years. Let's round it off and say he does it in 5 seconds each time. You've got a chart showing 5-5-5-5-5 going on for miles with Welker in the dash. Then, he tears a ligament in Houston. His time on his next weekday run is...maybe..30 seconds because he's walking it on crutches (but, it's Welker, so he's giving it his all on crutches). What's the best prediction for his Tuesday attempt? The median of the data would be 5. The average would be slightly above five depending on how far back it goes. But, the tiny sample size is the best prediction because he'll still be hurt. Something's changed...and the old sample size doesn't really matter any more until after surgery and rehab.

One could argue that Indy's approach is part of an evolutionary tick (don't want to suggest P-Manning is the only guy focused on moving chains with short passes, he seems to represent an extreme though). It may take emphasizing a smaller sample size, or changing the way people think about offense to fully capture the impact of what's going on. I still think changing our thought process out of a rushing/passing dichotomy into a moving the chains/attacking downfield dichotomy would better capture what's going on in pro football right now (as would looking at how defenses are dealing with those issues in reverse...some defenses allow the short stuff to protect against the long stuff...some attack aggressively accepting that they'll sometimes get burned for big plays).

To me, it's both a fault of DVOA and an error in interpretation. DVOA isn't fully capturing the impact of what Indy is doing...and the interpretation is seeing the running game as a "weakness" that needs "fixed."

40
by Jerry :: Tue, 02/23/2010 - 8:00pm

If there had been a Baseball Outsiders (and the supporting technology) in the early '80s, we'd have given Schmidt credit for "fixing his average" after '81, and then retracted that credit in the ensuing years. I haven't looked at several years of Colts numbers, but I'd need to see the trend play out before I reach a conclusion.

41
by Jeff Fogle :: Tue, 02/23/2010 - 9:40pm

I'd have to respectfully disagree with that Jerry. That's the exact opposite of how the sabermetric field has dealt with batting average, and outliers. And, Bill James was doing abstracts during the early to mid 80's, with Schmidt being one of his key examples for how the full scope of runs created isn't reflected well by batting average alone (he said something like, "the purpose of an offense is to create runs, not accumulate a high batting average."). Baseball Prospectus and the Abstract would consider the one year in batting average an outlier, while confirming Schimdt's role as one of the most dominant offensive forces of his era.

2009 for the Colts isn't out of line from the past couple of seasons (all I had time to look up). They've led the league in 3rd down conversions three years in a row. They're typically at or near the top in all the categories I listed earlier. They didn't have much of a running game in 2008 or 2009 (ranking last in rushing yardage two years ago). DVOA had them 2nd in passing offense back in 2008, but 27th in rushing. I'm sure it looked like they needed to "fix" their running game back then too...just before they went 16-1 this season in games that mattered despite NOT "fixing" it...then hearing again they need to "fix" rushing. They're making a choice. This year was a continuation of last year...and kind of the ultimate direction of philosophies developed in the Dungy years. That may help people concerned about sample size. Look at what happened in 2008 and 2009.

I should mention that DVOA has now undershot "estimated wins" eight years in a row for the Colts. And, this isn't pre-season estimates (not sure what those were, maybe somebody could look it up). This is once the stats are in the books...the way FO turns stats into records has undershot the actual record. I went back to look up the last three years at first so it would match the other stats I had looked up. Figured I'd go four years back just to see...and it just kept happening.

2009: 10.9 wins vs. a 14-2 eventual record
2008: 10.5 wins vs. a 12-4 eventual record
2007: 12.3 wins vs. a 13-3 eventual record
2006: 10.2 wins vs. a 12-4 eventual record
2005: 12.7 wins vs. a 14-2 eventual record
2004: 11.4 wins vs. a 12-4 eventual record
2003: 10.7 wins vs. a 12-4 eventual record
2002: 8.3 wins vs. a 10-6 eventual record

Tony Dungy took over the Colts in the 2002 season, and the current offense is still very much in that mold.

Maybe this past year shouldn't be termed an "evolutionary tick" as I called it earlier. More of a culmination of recent evolution...taking the risk/reward issue to its fullest by using short passes to do what the rushing game used to.

2009 and 2008 were very similar in terms of today's discussion. Use that as a sample size if there are concerns about just one year.

Maybe one of the intense math guys can calculate the odds of the projected record off the stats undershooting the actual record 8 years in a row.

42
by DeltaWhiskey :: Wed, 02/24/2010 - 4:03am

When I look at projected records, I arbitrarily give credit for being within +/- 1, but I do think this is a reasonable margin of error. I've looked at DVOA as a predictor of wins (simply using the correlation regression line for correlations wins w/ DVOA for the period from '94 through the '08 season). DVOA predicts wins within +/- 1 sixty-six percent (66%) of the time. The mean difference is zero (0) the standard deviation is 1.57. Therefore, for a team to have remarkably over/under performed their DVOA projection, they have to have missed by 3 (one could argue 4, which reduce the number to 2) or more games. Since 1994, only 27 teams have met the +/- 3 criteria, and only 12 have over-performed.

Using this methodology , the IND '02 to '09 data, the DVOA projection is correct 5 out of 8 (63%) of those years('08, '07, '05, '04, '03). Based on their 17.1% DVOA this year, IND would be expected to have won only 10 games, so this year's (2009)IND team significantly outperformed their DVOA. However, none of the other years cited are that far out.

43
by Jerry :: Wed, 02/24/2010 - 4:49am

I apologize for not expressing myself better above.

My point with Schmidt was that if one were to talk about Schmidt's "need to fix his batting average" after 1980, it would look like he had in 1981. It would take a while to realize that '81 was the outlier. (During that same while, James' writing would become popular enough that ignoring batting average could be discussed reasonably.)

I looked up the Colts rushing DVOA from 2009 back to 2002 (league rank in parens):

2009 -3.0 (22nd)
2008 -6.5 (27th)
2007 +11.9 (5th)
2006 +9.0 (7th)
2005 +8.7 (8th)
2004 +6.7 (11th)
2003 -5.5 (23rd)
2002 -3.9 (24th)

Maybe the last two years are an evolutionary step. Or maybe this is just a cycle. Or maybe something else entirely. Like Mike Schmidt's batting average, it will be easier to make that determination when looking back a few years down the road.

It looks like Estimated Wins are built primarily around Pythagorean Wins. In 2002, the Colts were 8.3 Est vs. 9.0 Pyth vs. 10 Actual. In 2006, they were 10.2 Est vs. 9.6 Pyth vs.12 Actual. In the other six years, Estimated and Pythagorean were within .3 of each other. As always, I'm willing to assume that Aaron hasn't missed some obvious fix; YMMV.

47
by Jeff Fogle :: Wed, 02/24/2010 - 5:50pm

Thanks again to both DW/Jerry for the dialogue.

First...DW...want to make sure I'm understanding. You're saying that you use a different method than FO does to estimate wins off DVOA...and that your different methodology is closer to estimated wins for the Colts in the years sited. So, even though FO's estimated wins is only within the reasonable threshold of +/-1 two times in the eight years sited, your regression technique is within that threshold five times in eight. Is that correct?

Does your regression still undershoot the Colts in all eight years? Or, do the Colts projections sometimes end up higher than reality? Would you not consider undershooting all eight years remarkable if that's the case...even if five of the eight years were within one win? Given the normal regression to the mean/plexiglass principle type stuff we see in sports, I would still think eight years of undershooting would be a clear unlikelihood.

I was thinking that turning the data into two-year hunks might better present the magnitude of the misses from the FO methodology. That way we're looking over 32-game hunks instead of 16...and allowing for the typical plexiglass stuff to show an influence (which is doesn't here but probably would for most teams I'm guessing).

08-09: 21.4 estimated versus 26 actual
07-08: 22.8 estimated versus 25 actual
06-07: 22.5 estimated versus 25 actual
05-06: 22.9 estimated versus 26 actual
04-05: 24.1 estimated versus 26 actual (first time within 2)
03-04: 22.1 estimated versus 24 actual
02-03: 19.0 estimated versus 22 actual

Given how often almost everything mathematical in sports shows regression over multi-year hunks...this really sticks out. Indy overachieves what their own stats say they should be doing every year in defiance of the whims of math...and that's been true through the Dungy era. Though, it sounds like it would stick out less with DW's preferred method.

Jerry, seems like we're mostly on the same page with the Schmidt stuff...though I'm concerned that FO is on the verge of looking like a 1981 baseball pundit rather than a 2010 "innovative stat" source with their thoughts on the Indy running game. As you show in your post:

Indy 22-27th in rushing dvoa: 48-16
Indy 5-11th in rushing dvoa: 51-13

Basically interchangeable, and even moreso percentage-wise if you don't count Dungy's first season as head coach. They averaged 12-4 in the four years they ranked in the 20's, 12.75 and 3.25 in the years they were in the top dozen. The rushing game isn't something that needs "fixed." They're moving the chains differently.

Went to Jim Armstrong's great drive stat page here at the FO site for some background info on the issue.

This past year, Indy was:
#1 in drive success rate
#2 in TD's per drive despite a poor rushing game that would theoretically have trouble finishing off drives
#30 in starting field position

Back in 2008, Indy was:
#1 in drive success rate
#3 in TD's per drive (same "despite")
#26 in starting field position

In 2007, when they had a much better running game in terms of DVOA
#2 in both (behind the Pats juggernaut), though field position was better at 12th.

In 2006, #1 in both despite being 27th in field position.

The rushing game has little connection to the end result for the Colts in terms of their drive success, or their success in the standings. I agree with you completely about the "maybe" element of all of this type of analysis. I do think the evidence is pretty strong here that DVOA is less reflective of the Colts offensive reality than Jim Armstrong's drive data is.

49
by DeltaWhiskey :: Thu, 02/25/2010 - 7:21am

"So, even though FO's estimated wins is only within the reasonable threshold of +/-1 two times in the eight years sited, your regression technique is within that threshold five times in eight. Is that correct?"

Yes, no, maybe.

My apologies for being confusing. I was referring primarily to the data you presented when I referenced '08, '07, '05, '04, '03 as being within one. I round off projected wins before subtracting from actual wins b/c winning fractions (except ties I guess) of games is not possible and wins are whole numbers. Just a preference. I've compared a few years of FO's projected wins to my projected win method and they usually don't differ much. FO's appears to be slightly more accurate. Therefore, when I threw out the mean difference of zero and Standard Deviation of 1.5, I was using my numbers. Sloppy, but easily available to me and probably pretty damn close to FO for this discussion.

Since you ask about my projections, here's how IND fairs using the straight up regression line for the DVOA:WINS correlation coefficient.

YEAR WINS PRED DIF
2003 12 11 1
2004 12 12 0
2005 14 13 1
2006 12 10 2
2007 13 12 1
2008 12 10 2
2009 14 10 4

"Does your regression still undershoot the Colts in all eight years? Or, do the Colts projections sometimes end up higher than reality? Would you not consider undershooting all eight years remarkable if that's the case...even if five of the eight years were within one win? Given the normal regression to the mean/plexiglass principle type stuff we see in sports, I would still think eight years of undershooting would be a clear unlikelihood."

With my methodology (b/c the data is convenient to me and easy to access), and looking at 2000 to 2009 b/c I started this w/ 2002 through 2009, but decided a decade review might be more interesting) I found the following:

In the chart - NEG refers to teams that undershot their projected wins, ZERO refers to teams that hit exactly their projected wins, and POS refers to teams that overperformed. This table does not utilize my preferred +/- 1 margin.

TM NEG ZERO POS
ARI 1 2 7
ATL 0 4 6
BAL 6 1 3
BUF 5 4 1
CAR 4 0 6
CHI 2 2 6
CIN 5 2 3
CLE 4 4 2
DAL 3 5 2
DEN 4 3 3
DET 4 2 4
GB 4 2 4
HOU 1 3 6
IND 1 2 7
JAC 6 2 2
KC 7 2 1
MIA 5 2 3
MIN 0 2 8
NE 3 1 6
NO 2 3 5
NYG 3 3 4
NYJ 6 3 1
OAK 6 2 2
PHI 6 0 4
PIT 6 1 3
SD 5 1 4
SEA 2 3 5
SF 2 3 5
STL 1 3 6
TB 8 1 1
TEN 2 2 6
WAS 7 3 0
AVG 3.7 2.2 3.9
SD 2.2 1.1 2.1

What these results suggest is, IND did indeed fairly consistently outperform their DVOA as did two or three other teams. ARI and MIN did as well or better, and arguably, ATL performed in a similar manner. Now the difference in IND and these other 2 (3) teams is the level at which they performed over the 10 year span. ARI managed 2 playoff appearances and MIN appeared 4 times (ATL = 3), while IND appeared 9 times.

So...fans who should be heartbroken about the last decade, as their team left some wins on the table more than 50% of the time - BAL, BUF, CIN, JAC, KC, MIA, NYJ, OAK, PHI*, PIT, SD, TB, WAS. Fans who should be thankful b/c their team stole wins more than 50% of the time - ARI, ATL, CAR, CHI, HOU, IND, MIN, NE, NO, SEA, SF, STL, TEN.

*See, DVOA does love PHI (but not as much as it loves WAS, KC and TB)

50
by Jeff Fogle :: Thu, 02/25/2010 - 4:41pm

Thanks DW. Very interesting to read through your data. Appreciate you taking the time to put that together. My eye was immediately drawn to New England because I always link Indy and NE in terms of how they drive the field so effectively.

You show:
Indy: 1-2-7
NE: 3-1-6

That's a combined 4-3-13...suggesting a strong (to me at least) relationship between success at driving the field and overachieving projections. Am I right that the 4-3-13 over a combined 20 years was relatively frontloaded with the NEG numbers, but backloaded with the POS numbers during these recent years when Manning and Brady have been such a big part of public consciousness? It would look that way based on the FO estimated wins pages. 2009 was the first time in a while NE didn't get past the estimation...and it's a year Brady was dealing with myriad injuries (and teammate injuries)

2009: 11.2 estimated wins, 10-6 actual record
2008: 9.6 estimated wins, 11-5 actual record (without Brady!)
2007: 14.3 estimated wins, 16-0 actual record
2006: 11.6 estimated wins, 12-4 actual record (even if you round off)
2005: 9.7 estimated wins, 10-6 actual record (even if you round off)
2004: 13.1 estimated wins, 14-2 actual record
2003: 11.5 estimated wins, 14-2 actual record

Closer than with the Colts, but still only one example on the list where the actual record was BENEATH the estimation...and that was last year with Brady being banged up. The 2008 mark suggests something systemic since Brady was out...meaning the Belichick approach (and the talent he acquired) played a big influence too obviously. Don't want to suggest it's just a Manning/Brady thing. Looks like a combination of coaching philosophy/schematic/great QB/talented weaponry stuff in play. The point being...the FO method for estimating wins may not fully capture the potential of the most lethal combo's.

And, I guess that's kind of the heart of the debate about the Philly issue. They've been underachieving four of the last five years

2009: 10.9 estimated wins, 11-5 actual record (0.1 away from even)
2008: 11.7 estimated wins, 9-6-1 actual record
2007: 10.2 estimated wins, 8-8 actual record
2006: 11.9 estimated wins, 10-6 actual record
2005: 7.7 estimated wins, 6-10 actual record

Philly had a great run before that where they overachieved in '04-'02. Can see why some think the small sample sizes create illusions in that light. For someone like me, who's used to seeing the regression stuff happen from year to year rather than in five year hunks, it's at least odd that Indy and NE have been so consistently better than their stat estimations during the heart of the Manning/Brady years (except for NE last year), while Philly has been +0.1, -2.7, -2.2, -1.9, and -1.7 the last five years.

Thanks again for putting all that together DW...

53
by DeltaWhiskey :: Fri, 02/26/2010 - 8:23am

"That's a combined 4-3-13...suggesting a strong (to me at least) relationship between success at driving the field and overachieving projections"

Time to put this to the test. Below are correlations of the difference between Expected Wins and various FO Drive Stats for the period 2000-2010. Expected wins are calculated using the same equation as I've used before (DIF = difference betweene expected wins and actual).

The first column are the variables compared, the second the correlation coefficient, the third the p value, and the last number is the R-Square

Variables
DIF:YDS/DR 0.124 p < 0.01 R-Square 0.016
DIF:PTS/DR 0.178 p < 0.01 R-Square 0.032
DIF:TDS/DR 0.155 p < 0.01 R-Square 0.024
DIF:DSR 0.103 p < 0.05 R-Square 0.011

The statistically significant results suggest a relationship between Drive Efficiency and expected performance exists; however, the small correlation and resulting low R-square suggest it is an incredibly small relation and that other factors seem to be more important and relevant.

I conducted a second analyis of only those teams that had exceeded expectations by one or more games, this subset of data yielded nonsignificant results, as did an additional analysis of this subset of data that consisted only of teams that had won 9 or more games.

However, slicing the data pie in a different manner yielded the following:

Again, The first column are the variables compared, the second the correlation coefficient, the third the p value, and the last number is the R-Square.

Variables r
DIF:YDS/DR 0.553 p < 0.01 R-Square 0.306
DIF:PTS/DR 0.572 p < 0.01 R-Square 0.327
DIF:TDS/DR 0.429 p < 0.01 R-Square 0.184
DIF:DSR 0.522 p < 0.01 R-Square 0.272

For this analysis, I looked only at the more efficient teams in terms of drive performance. To select these teams, I only analyzed those that were at least one SD above the mean for the particular drive measure being assessed. As seen above, their does seem to be a relationship between drive efficiency at the higher end and exceeding expected wins. That is, teams that are efficient at driving the ball are at least in the high average to above average range are likely to exceed their expected wins.

54
by Jeff Fogle :: Sat, 02/27/2010 - 12:09am

Thanks DW. Glad your test seems to be in line with the theory about drive efficiency helping teams exceed expected wins. Any conclusions we may be able to draw from that? Or, at least hypothesis we can make?

Between hearing a lot of talk about football teams trying to squeeze out an extra possession in halves (Mack Brown talks about it down here in Austin), watching NBA teams try to go 2 for 1 with possessions in the final minute of quarters, and even seeing Olympic curling where teams try to position themselves to have the hammer in the 10th end...I'm thinking there may be something here about the very best drive teams having an edge late in close games if they can get themselves a crunch time possession with time on the clock. I know it's kind of a stathead blasphemy to talk about "knowing" how to win close games...we may be coming upon something that actually creates that based on what Indy and NE have been doing now over a measurable time frame going back a few years. At least I'm taking that as a possibility to take very seriously based on your research.

55
by DeltaWhiskey :: Mon, 03/01/2010 - 10:20am

I think it's early for strong statements and hypotheses. I think we can say that one way to Win more than would be expected (i.e. Predicted by DVOA) it to have high drive efficiency.
I did some follow-up analyses comparing the mean number of wins for the above average efficiency teams to the average to below average efficiency teams in the sample of teams that exceeded DVOA predicted wins by at least one. The above average efficiency teams averaged 11.3 wins while the average to below average teams averaged 9.08 wins. There was no statistically significant difference between the magnitudes of the difference by which they exceeded their predictions. On average they exceeded their DVOA predicted wins by 1.6.

Based on the above, I can tenatively state that if a team is above average in drive efficiency, between 27% and 43% of the variance in the degree of the diffence by which the team exceeds its DVOA predicted wins is due to being above average in drive efficiency.

I think the next step for me is to look at this from the standpoint of defensive efficiency and to look closer at the cellar dwellers, both in drive efficiency and failure to meet DVOA predicted wins.

"I know it's kind of a stathead blasphemy to talk about "knowing" how to win close games...we may be coming upon something that actually creates that based on what Indy and NE have been doing now over a measurable time frame going back a few years"

I'm going to have to think abour this a little, never underestimate the value of "luck" in exceeding DVOA predicted wins.

51
by commissionerleaf :: Thu, 02/25/2010 - 8:32pm

I happen to disagree.

The Colts have not gone to a short passing game because it's less risky or more reliable than a long passing game or a strong running attack. The Colts offense was far more reliable and potent in both 2004 (passing) and 2005 (running) than in 2009. The Colts have gone to a short passing attack almost exclusively because they no longer have an offensive line sufficient to do anything else. They don't have a choice.

Because Manning has such a quick release and ability to get plays called at the line, people don't realize just how awful the Colts' offensive line is now. Charley Johnson, who wasn't good enough to back up Tarik Glenn once upon a time, is the starting left tackle. Ryan Lilja didn't play in '08 and was not the same player in '09. Ryan Diem is a good player, but not a dominant right tackle. Kyle Devan... is a decent backup Center. The Tony Ugoh drafting may eventually be seen as the beginning of the end for the Colts... a Robert Gallery level mistake. Of course, Manning's ability to transition to a West Coast-ish system rather than passing downfield and handing off has prevented the decline from being as visible.

The rule changes and enforcement changes favoring passing in general have helped too, by masking the downward relative trend in Manning's numbers.

52
by Jeff Fogle :: Thu, 02/25/2010 - 10:13pm

Can't disagree about the quality of the line given the various metrics out there. Only their decision-makers know for sure how much is a choice in style, and how much is from a lack of choice. I think it's tough to talk about a decline though when we're talking about the league MVP playing on a team that went 16-1 in the games they were trying to win. They're still doing a better job than anybody of driving the field for TD/success based on the drive data here at the FO site.

2004 was truly a monster year. I think we have to be careful assuming that's some sort of repeatable peak. It was a big outlier in terms of TD passes. Manning tied his second best TD performance this year. He had his third highest yards-per-attempt this year even though they emphasized shorter stuff. Everyone looks worse than their peak year most of the time, doesn't mean it represents a significant "decline." More like a regression to the mean off the outlier.

I guess I should say too that the Colts DID win rushing yardage in the Super Bowl 99-51, and over the Jets 101-86. That 461 total yardage effort vs. NYJ would seem out of character for a team with an "awful" offensive line. Obviously an outlier too from both sides of that coin. I think the author said Addai was fine, but it was the backups who weren't doing enough (going from memory because I commented off your comment rather than the original article here). If Manning wins an MVP, and Addai had respectable numbers (before being a weapon in the playoffs), it's tough to look too harshly at the line I think. Unless Manning would be a super-duperstar and Addai an MVP candidate with a normal line or something. Tough to add up to elite production drive-wise with an "awful" line.

22
by mm (not verified) :: Tue, 02/23/2010 - 1:33am

The Texans should look to the Super Bowl champions to fill the biggest question mark on their roster. The Saints will likely retain restricted free agent running back Pierre Thomas, but Mike Bell and his 54 percent Success Rate (sixth in the league) could be available.

I'm pretty sure the Texans had Mike Bell in their training camp before the 2008 season; he pulled a hamstring in the first few days and was cut soon after, leaving him out of a job until the Saints signed him later that year.

So while the Saints may lose him, the Texans will likely not be his first choice.

26
by Theo :: Tue, 02/23/2010 - 5:09am

"The fans in Jacksonville might prefer hometown hero Tim Tebow..."

No, they want safety, offensive line, receiver, real QB.

27
by lobolafcadio :: Tue, 02/23/2010 - 5:36am

Or just best available player... Please, no draft for need, it gave us Reggie Williams, Marcedes Lewis, Reggie Nelson, Derrick Harvey (Matt Jones was a gamble, to make it clear a bad gamble).
You're just one bad play away from a need at what seemed a very good position on your roster.

36
by Eddo :: Tue, 02/23/2010 - 2:22pm

Agreed. It's not "fans in Jacksonville" that prefer Tebow, it's "fans in northern Florida that care much more about college football than the NFL".

31
by Tim in London (not verified) :: Tue, 02/23/2010 - 9:56am

Seems that the Jags re-upped for one year on Spencer TMQ's boo boo Ernest Wilford

33
by bingo762 :: Tue, 02/23/2010 - 11:51am

Next stop, Philadelphia

34
by Sophandros :: Tue, 02/23/2010 - 12:25pm

"...Arian Foster would have finished second in the league in DVOA if he had carried the ball often enough to qualify."

Interesting statement.
-------------
Sports talk radio and sports message boards are the killing fields of intellectual discourse.

44
by Tom Gower :: Wed, 02/24/2010 - 10:57am

For those interested, I have expanded thoughts on Pisa Tinoisamoa as a Titans target on the Titans site I write for.