In this week's Varsity Numbers, Bill Connelly revisits some measures and concepts: Adjusted Scores, Covariance, and momentum (or whatever you choose to call it).
18 Nov 2006
Guest Column by David Lewin
With Jason Campbell set to start this week I'm sure many FO readers have gone scrambling for their copy of PFP 2006 to see what the college to pros quarterback projection system said about him. If you haven't looked it up yet let me save you the time: Jason Campbell is going to be good.
To refresh everyone's memory, the system is based primarily on games started and completion percentage. It gives a career value for a player which, using the typical career progression of quarterbacks, can be used to estimate performance in any given year. So, in honor of Jason Campbell's first start let's recap how the system has done so far this year and what we can expect the rest of the way.
Remember that the system is based on first- and second-round quarterbacks, which is why you won't see anything about Bruce Gradkowski or other late-round picks. Also, while the article in the book focused on peak projections, the system does project each player's early years, and those projections are referenced here.
Also known as the poster boy for the projection system. I actually suggested putting Rivers on the cover of PFP 2006, but instead they went with two New York players on the cover to promote the book in the biggest market.
(Ed. Note: Last time I make that mistake. Rivers will be on the cover of next year's book.)
Although Rivers had the highest projection of any player even I wasn't sure he would be this good this fast. The projection for Rivers this year was 5.8 DPAR per game, which would have been good for fourth in the league last year. He has exceeded even his lofty projection, averaging 7.4 DPAR per game. I personally haven't had the opportunity to see Rivers play very much this year, so I don't know how he has done this, but I would not be surprised to see him continue this remarkable level of success and become the best quarterback of the vaunted 2004 class (Eli Manning, Rivers, Ben Roethlisberger, and J.P. Losman).
Grossman has been up and down this year. He has played brilliantly at times, and stunk it up mightily at other times. He has shown the ability to throw one of the prettiest deep balls in the game, but he seems unable to throw to the sidelines especially on out patterns. Many people have mentioned Grossman's erratic mechanics, and that is probably part of the problem. In order to throw the ball to the sidelines effectively a quarterback has to open his hips and point his front foot to his target. Grossman has been inconsistent with doing this, and has often thrown off his back foot or across his body. When throwing downfield your feet are usually already more or less lined up, so stepping towards your target is less of an issue. If Grossman can develop into more than just a great deep ball thrower he might become a very good player in this league. As it is he is playing right along with his projection to develop into â€œa quality NFL starter,â€? (quoting myself in PFP 2006) and is virtually matching his projection of 2.5 DPAR per game for this year.
After a rookie year in which Smith was worse than even Ryan Leaf, I can't say that I really expected Smith to bounce back to his above-average projection. He hasn't bounced all the way back to it yet (he was projected slightly above 2.0 DPAR per game this year and he's right around 1.5) but he has shown significant promise. One improvement I am working on making to the projection system for next year is the inclusion of age as a variable. Right now, only number of years in NFL is included. In the age of redshirts (and holding kids back in grade school for athletic reasons; I'm looking at you Clausen family) most quarterbacks enter the league at age 23 or 24. Smith, however, was 21 as a rookie, probably a disadvantage when it comes to his early performance.
Like Grossman, Leinart has had an up and down season. He has shown flashes of brilliance, like the first half of the Chicago game, but also surprisingly lengthy periods of ineptitude. The Cardinals have some offensive weapons, but an inability to run the ball is clearly hindering the team. It seems they might have a good idea to have taken some of the 20 million plus they have left under the cap and make a run at Steve Hutchinson, but these are the Arizona Cardinals, so they didn't. I still think that in the long run Matt Leinart will be star, and he is not that far off his projection right now (he's averaging 1.0 DPAR per game, he was projected to be at 2.4). A lot will depend on whether the Cardinals choose to put a competent offensive line up in front of him.
One of the biggest surprises of the season in my mind has been the play of Vince Young. He is much farther along as a passer than anyone could have been anticipated (although his DPAR per game of 0.0 is actually below his projection of 0.4), but his running has been much less of a factor than I would have thought. In fact, his running has been worth -5.8 DPAR this season. This cannot continue to be true for a runner as talented as Young. He has clearly demonstrated talent as a passer, but the most important thing to his development will be whether he can effectively learn when to utilize his running ability to complement his passing.
As I wrote in PFP 2006, "The Bills are really going to regret passing on Matt Leinart." Thus far this has been true, but more because Losman, who continues to underperform his less than optimistic projection, has been that bad than because Leinart has been good. I don't see Losman having much of a future in the league past this year, so maybe the Bills will be in the market for a real prospect like Brady Quinn come April.
Eli Manning will never be Peyton, but he's pretty good; Byron Leftwich deserves to be a starting quarterback somewhere; David Carr is not terrible; and Ben Roethlisberger should bounce back eventually as long as he avoids sky-diving without a parachute and similar activities.
And now on to the alleged subject of this column...
I don't know a single Redskins fan not excited for the Jason Campbell era to start. According to FO's advanced metrics, Brunell was not playing exceptionally poorly (ninth in DVOA), but anyone watching the Redskins play this year could see that even if the end is not here yet for Brunell, it is definitely in sight. A distinct inability to throw the ball downfield to gamebreaking receivers like Santana Moss and Antwaan Randle El has really limited the Redskins offense. With Campbell at the helm, expect that to change in a hurry.
If there is one way to describe Campbell as a quarterback it would be prototypical. He is 6'4", 230 pounds with good mobility and the cannon arm that everyone loves. His college stats profile of 39 starts and 64.6% completion rate is quite favorable as well.
However, the devil is in the details. Campbell has played in six offensive systems in the past six years. I have seen him described a slow learner, but this is probably related to him having to learn a new system every year. I have to imagine that Al Saunders would not have given Campbell the keys to this offense and its 700-page manual (playbook) if he wasn't comfortable with it.
Campbell's college stats also have on major red flag, his number of attempts. In general I don't use attempts in the projection formula because number of attempts tends to be highly collinear with games started. In the case of Campbell, because he played with Ronnie Brown, Cadillac Williams, and Brandon Jacobs at Auburn, he was not asked to throw the ball very often. If I used attempts instead of games started the formula is, in general, almost as accurate, but would give a much lower projection for Campbell.
I suspect this means that Campbell will slightly underperform his projection, which for this year is about 4.0 DPAR per game. Still, I expect Campbell to be an immediate improvement over Brunell, despite a decent number of rookie mistakes. By the end of the year, Redskins should be looking forward to a bright future, at least at the quarterback position.
David J. Lewin is a 19 year old sophomore at Macalester College where he plays football. He is currently trying to round up job offers for the summer, some of which will hopefully involve his alleged "skills" as a sports statistics analyst. He can be reached at dlew33-at-yahoo.com. To read more about the quarterback projection system, read this article from FO or check out the article on page 433 of Pro Football Prospectus 2006. Look for his column evaluating the quarterback prospects for the 2007 NFL draft sometime this off-season.
48 comments, Last at 29 Nov 2006, 10:27pm by David Lewin