The Football Outsiders Almanac came out recently, and Bill Barnwell, who is one of the authors of the book, was nice enough to answer some questions about the FO methods and the Titans. The intelligent questions were submitted by Hal, and the unintelligent ones were submitted by me.
Is there a problem with strategic expectation bias? The system goes to great lengths to normalize performance based upon the quality of opponent, but one hole I see is that from week to week a given opponent inserts different game planning based up the strengths of the opponent. This may not be a problem for teams that are somewhat balanced in terms of strengths andweaknesses, but if a team has a great strength anda glaring weakness, it will fundamentally alter the way the opponent approaches the game. One example from last year might be Sammy Morris from the Patriots. No one with a straight face is going to claim he is somewhere between the 8th and 5th best running back in the league, as his DYAR and DVOAresults suggest. But New Englandwasn't seeing any 8 man boxes or run blitzes like teams with imposing running games. Shouldn't the "defense adjustment" portion of the formula be able to recognize that the defense as it related to New England's run blocking (ranked 3rd) and Sammy Morris was incredibly weak because they were so worried about Moss and Welker?
That's quite the question! I think there's something to be said for any statistic in not totally analyzing every possible twist and turn with regards to how a team or a player is viewed, and certainly, DVOA is no exception.
The issue with the sort of adjustment you're suggesting is that it's not really quantifiable -- you and I know from watching games that it's easier for Morris to get a running lane than it would be for, say, Steven Jackson, but we can't exactly plug in that Morris is only seeing an average of 6.3 guys in the box before each snap.
That's why I wouldn't sit here andtell you that Sammy Morris is, in a vacuum, the fifth-best running back in the league. I'd tell you that given the opportunities he gets in the New Englandoffense with Tom Brady at quarterback, the Patriots offensive line blocking for him, and Randy Moss and Wes Welkeron the outside, he's the fifth-best running back in the league.
Of course, every other football statistic is subject to the same vagaries. I don't believe in reducing players down to a metric or a number, even though I believe that DVOA is a useful tool. You should use it, especially when judging the individual performance of players, as a statistical tool in combination with what your eyes tell you.
(small point of correction: the Titans DID make an outside move to replace Haynesworth, they added DT Jovan Haye from Tampa this off season)
That's true, we forgot to mention Haye in the chapter. I actually like Haye as a bit of a sleeper prospect; he's agile enough on the interior that he could become a worthwhile player in a rotation. I still think Jason Jones is the future there, though.
Your section on the Titans predictably goes into depth on the issue of replacing Albert Haynesworth. It mentions how there is no precedent for losing an All-Pro defensive tackle the year after. However, Albert missed 10 games over the last 3 years. Could you look at the effect his absence had on the Titans' defensive line performance based upon those games?
Sure thing. The following is the Titans' DVOA with and without Haynesworth split by runs and passes over each of the last three seasons. Remember that DVOA measures performance versus the league average, so for defense, a negative number is preferable to a positive one.
With Haynesworth: 5.3% Pass DVOA, 9.0% Run DVOA
Without Haynesworth: 3.4% Pass DVOA, 0.2% Run DVOA
With Haynesworth: -29.6% Pass DVOA, -9.9% Run DVOA
Without Haynesworth: -38.1% Pass DVOA, 10.1% Run DVOA
With Haynesworth: -21.8% Pass DVOA, -3.8% Run DVOA
Without Haynesworth: -11.8% Pass DVOA, -22.1% Run DVOA
The 2007 numbers really stand out as the starkest, andmatch up withwhat I remember from watching the Titans play that year; they looked like a totally different team with Haynesworth out of the lineup. In 2008, that wasn't as pronounced, but the sample was a game smaller and came after they'd already clinched home field advantage. I think having Jones andJacob Ford around also made their line look better.
The Titans plan to go with a platoon strategy on the D-line, and a few key players notwithstanding, they always have been a unit that plays a rotation. Do you have any data on platoon strategies vs. fixed first string approaches which hesitate more to use 2nd string players?
Not comprehensively. Anecdotally, we findthat teams like the Redskins that build around a stars andscrubs approach tend to fail -- injuries are just too pervasive in the league to build that sort of team. If you look at the teams that have won the Super Bowl the past few years -- the Steelers, Colts, and Giants -- each rank among the best-drafting teams in the league, getting significant contributions up and down the lineup from their full roster. I suspect that the Titans might be better off having a defensive line that goes eight deep than one that just plays the same four guys over and over again until they get hurt, even if it's players of superior talent.
Your preview mentions that the Titans'offense will have difficulty making the jump from "good" to "great". I would contend we had the worst receiving corps in the NFL last year, and that fact is why the running attack only ranked 8th instead of being top 3. We kept the creme of that crop, Justin Gage, and added Washington, Britt and Jared Cook (pure receiving tight end). If that bumps our receiving group from "abysmal" to "serviceable", couldn't that be enough to force other teams to stay honest and therefore make things easier for the running game while improving the passing numbers? Doesn't the fact that we had plenty of room to improve in the downfield passing game defeat the logic that the jump from "good" to "great" will be very difficult?
The thing with the sort of statement you make above is that it's very subjective and can be spun a lot of ways. I could suggest that Gage was a mediocre receiver despite having a great running game to open up holes for him, Washington has been one of the worst receivers in the league at dropping passes and consistently has performed below the level of his teammates, that Britt has looked awful in camp, and that rookie receivers -- especially rookie tight ends -- often fail to make an impact, even when they're at the Calvin Johnson level, let alone the Kenny Britt/Jared Cook level.
Would that analysis be fair? Not really, even though like your statement, there's some truth to the logic behind it.
What we've found in our research is that the dominant factor in determining the effectiveness and reliability of a passing game is the quarterback. And while Kerry Collins is serviceable, I think you'd really have to be drinking the Kool-Aid to call him a very good quarterback.
Another, much more important factor is that the Titans had an extremely healthy offense last year. Their offensive linemen made it to 79 of 80 games, with Kevin Mawae missing a meaningless Week 17 game. Justin Gage and Justin McCareins combined for six missed games, Paul Williams and Quinton Ganther were "injured" for a few games, and Vince Young was hurt for three weeks. That's really not very much at all -- in fact, the Titans were the third-healthiest offense in football last year.
That's not likely to reoccur in 2009; just because of the nature of the game, teams get hurt more frequently than that. You saw what happened to the Titans offense in the second half of the Ravens game after Chris Johnson was hurt; an injury to Johnson or someone like Michael Roos would have a far greater negative impact on the team than the potential improvements made at receiver this offseason will have.
You've studied the Titans numbers and broken down the film, what "hidden facts" emerge out of the 2008 Titans that the average fan might not have seen on TV?
The lack of injuries is a big one; it's not necessarily that fans don't see that, but just that it doesn't really occur to you unless you really look at the broader picture. It's easy to think about injuries when your team is decimated by them, but the opposite isn't always true; you don't see, for example, Dolphins fans saying that they had a great year because they were the healthiest team in the league, even though lack of injuries has a very strong relationship with winning.
Beyond that? There are a lot of little things we see and pick up on from breaking down tape and putting together statistics. The Titans were way better on play action with Kerry Collins, going from 22nd in the league on play action to eighth. They ran way more plays, going from 26th in the league in offensive pace to second. The offensive line blew only seven blocks, tied for #1 in the league.
What do you think of the Titans 2009 draft class?
I'm not sure yet, to be honest. I'm not a big fan of taking wide receivers at the end of the first round, as they tend not to be great values; the following guys have been taken within five picks of the 30 slot this decade:
R. Jay Soward, Dennis Northcutt, Freddie Mitchell, Reggie Wayne, Quincy Morgan, Jabar Gaffney, Michael Jenkins, Rashaun Woods, Roddy White, Reggie Brown, Santonio Holmes, Robert Meachem, Craig Davis, Anthony Gonzalez, Donnie Avery, Devin Thomas
That's essentially two stars, two guys about to break out on the league (Holmes and Gonzalez), a few players who need a couple more years to figure out one way or another, and a whole bunch of busts. Throw in this organization's issues with identifying and acquiring wide receivers in the draft in the past and, well, I don't have high hopes.
I like the Sen'Derrick Marks pick, though. I think he's a very good fit for them schematically, and he looked good in the Hall of Fame game once I figured out he was #94.
We all think that Michael Griffin is one of the top safeties in the league. Does your analysis of him back that up?
I would say he looks pretty good so far. His pass defense metrics were all right around 15th or so in the league last year amongst safeties, which is a good sign. I think he benefits from having such a great front seven in front of him, but he's certainly holding his own. He might not be at the Reed/Polamalu level yet, but he's right below that level.