DVOA?...

[Mongo3451]

http://blogs.suntimes.com/bears/2009/07/19...ball_outsi.html

 

For all you stat guys out there. I like the way these guys look at stats. It has a link within the article for the entire book if anyone is interested. I'm going to start trying to digest it. It looks as though it may be a good FF tool as well. They are going to post more articles on this later.

 

 

[defiantgiant]

http://blogs.suntimes.com/bears/2009/07/19...ball_outsi.html

 

For all you stat guys out there. I like the way these guys look at stats. It has a link within the article for the entire book if anyone is interested. I'm going to start trying to digest it. It looks as though it may be a good FF tool as well. They are going to post more articles on this later.

 

DVOA has some value as a statistic, but should be taken with a large grain of salt. As an example, if you only look at DVOA ratings, you would have to think that Matt Forte is actually a significantly worse-than-average running back. His negative 7.1% DVOA is 38th in the league. Why is this? Because, by FO's own admission, DVOA and their other statistics, "do not separate the performance of a running back from the performance of his offensive line."

 

If you go to their page on Adjusted Line Yards (a stat that I much prefer to DVOA) you can see that the Bears' line ranked 25th in the league in ALY. No team ranked 25th or worse in ALY had a back in the top 20 for DVOA, so I think it's safe to say that a back's DVOA can be skewed by a poor offensive line.

 

Furthermore, according to FO, "a team with a low ranking in Adjusted Line Yards but a high ranking in [carries of] 10+ Yards is heavily dependent on its running back breaking long runs to make the running game work." The Bears ranked 25th in ALY but 16th in carries of at least 10 yards, so I think it's fair to say that it was the offensive line holding Forte back, not the other way around. Still, if you only look at his DVOA, you'd think he was a pretty lousy running back.

 

I'm not saying DVOA is a bad stat, but it needs to be taken in context, just like YPC or any other stat.

[defiantgiant]

More on problematizing DVOA:

 

Just like a running back's DVOA is inseparable from his offensive line's performance, a wide receiver's DVOA is inseparable from his QB's performance. According to Football Outsiders, "we cannot yet fully separate the performance of a receiver from the performance of his quarterback. Be aware that one will affect the other."

 

If you believe DVOA alone, Dwayne Bowe is the 66th best wide receiver in the league, worse than the Jaguars' Reggie Williams. But Bowe is an excellent wide receiver with a bad quarterback, whereas Williams is a bad receiver with a pretty good quarterback.

 

It makes sense that all the Bears' receivers had negative DVOA: even with a good quarterback, they may have been in the red. But Greg Olsen has a DVOA of 0.0, and he's far from a league-average receiving TE (note: DVOA only measures tight ends as receivers.) Again, if you take his DVOA out of context, you'd think that Kevin Boss and Brent Celek are better receiving TEs than Olsen, when they're actually less talented receivers with more talented quarterbacks. Likewise, you'd think that the Broncos and the Dolphins each have TWO tight ends better than Olsen, and that both the Panthers' TEs (Dante Rosario and Jeff King, whose name I had to look up) are better than Kellen freaking Winslow, who's arguably one of the top 5 receiving TEs in the NFL.

 

Are Dante Rosario and Jeff King really better receiving TEs than Kellen Winslow? No way, but they do have a significantly better QB. The same is true of the Broncos and Dolphins last season: both Jay Cutler and Chad Pennington were more effective than Kyle Orton, and it's reflected in the DVOA of those teams' tight ends.

 

EDIT:

I should say this: as much as I have a problem with DVOA as an indicator of individual performance, I do have to admit that Football Outsiders' individual projections for the Bears' players are pretty close to what I'm expecting:

 

Hester: 62 receptions, 858 yards (13.8 avg.) and 5 touchdowns. It adds up to a 2.7 percent DVOA.

 

Olsen: 61 receptions, 738 yards, 7 touchdowns, 17.0 percent DVOA

 

Matt Forte: 322 rushes, 1,272 yards, 4.0 average, 68 receptions, 417 yards, 14 touchdowns, 2 touchdowns, 8.4 percent DVOA (as a rusher)

 

Cutler: 301 completions, 493 attempts, 61.2 percent, 3,409 yards, 20 touchdowns, 12 interceptions, 3.0 percent DVOA.

 

I think Cutler and Hester could both do a little better than they're projecting, but I wouldn't be mad at all if their projections were accurate.

[nfoligno]

Comments on the individual stat projections.

 

Hester: 62 receptions, 858 yards (13.8 avg.) and 5 touchdowns. It adds up to a 2.7 percent DVOA

 

The biggest issue I have here is the YPC avg. I think most would agree Hester's vertical potential was limited last year. Last year, we had a QB/OL and system which simply did not allow for great vertical opportunities, and even when they were there, Orton too often failed to connect. W/ all those limitations, he had a 13 ypc avg. This year, we have one of the strongest arms, and most aggressive mentality QBs in the game. Further, last year we had an OL so bad the OC himself said we limited the QB to 3 step drops to limit pass pressure. In such a system, you simply are not going to have many downfield opportunities. This year, we have made a lot of moves to upgrade the OL. Yet w/ these upgrades, they are only projecting a .8 increase in Hester's ypc avg. That is what I have an issue w/. I just feel, as much as w/ any single stat, his ypc avg will see a sizable increase.

 

I think Hester very well could have a ypc avg closer than not to 15.

 

Olsen: 61 receptions, 738 yards, 7 touchdowns, 17.0 percent DVOA

 

These are not bad numbers, and represent an increase across the board. One issue I have w/ their report is they mention how Olsen can be a nice possession receiver/target. While I am not saying he can not be that, I think many would argue he can be so much more. Clark is a TE I would label as being more of a possession receiving target. Olsen has been used as a possession guy, but the OL has been the top reason for that, not Olsen himself, who has downfield skills. No major issue w/ the stats, but just w/ the idea that Olsen is a possession receiver/TE.

 

Matt Forte: 322 rushes, 1,272 yards, 4.0 average, 68 receptions, 417 yards, 14 touchdowns, 2 touchdowns, 8.4 percent DVOA (as a rusher)

 

If Forte is on my FF roster, I would love these stats, especially as I usually play in PPC leagues. On the other hand, I am not sure I would be thrilled, as a bear fan, if this is what Forte does this year.

 

(a) 322 rushes? We have talked about this quite a bit of late. While I am not sure his number of carries goes down a ton, at the same time, 322 actually represents an increase. For Forte to get 322 carries, I think it would most likely point to K.Jones not producing much, thus needing to use Forte more, again.

 

( 4 ypc avg? 4 ypc isn't "bad", but I really think most here have greater expectations. He had a 3.8 ypc avg behins a dreadful OL, and w/ a QB that created a lot of stacked boxes due to his lack of ability to go downfield. The common belief is the addition of Cutler spreads out a defense more and the upgrade to the OL creates more holes. Thus, I think most expect a higher ypc avg for Forte. For him to have only a 4 ypc avg (.2 greater than last year), I think either (a) Cutler isn't effective enough to spread the field and/or ( the OL is not playing to the level of expectation after offseason upgrades.

 

© 68 catches? Forte is a solid pass receiving RB, as shown last year w/ his 63 catches. However, did we pass to him so much last year because we wanted to, or because we had to. For him to catch 68 this year, IMHO, it would point to multiple issues like (a) OL not playing well and Cutler having to dump off more often, ( WRs not stepping up and Cutler having to use his checkdown more. If I have him on my FF roster, I would love 68 catches, but as a bear fan, I would seriously worry about him having so many catches.

 

Further, per our coaches, one of the key areas we are hoping to improve w/ Forte is his yp catch. Last year his ypc was 7.6, which the staff felt was low and a key area in need of improvement. The projected numbers actually has his ypc dropping to 6.1. As w/ other comment, this to me would point to a poor OL and the majority of his catches being on jumpoffs.

 

Cutler: 301 completions, 493 attempts, 61.2 percent, 3,409 yards, 20 touchdowns, 12 interceptions, 3.0 percent DVOA.

 

I don't necessarily have an issue here. I believe his stats will be significantly better, but I also understand the reasons to project on the conservative side. The only issue I guess I have is, while projecting a more conservative offense and thus, stats, they also are projecting a decrease (slight) in his completion percentage. I would think if we threw less often and were more conservative, his completion percentage would go up.

 

When you look at the catches/yardage stats by themselves, they don't appear too bad, but when you look a little deeper, I think these stats look pretty bad when you consider our expectations, not simply statistical, but quality. To me, much of these projected stats woudl indicate our OL not playing very well, and our offense not really being as overall solid as expected.