Writers of Pro Football Prospectus 2008

06 Oct 2009

Week 4 Quick Reads

by Bill Barnwell

Football can be a confusing sport at times, even for the educated viewer. Anything beyond the most cursory analysis of what went right (or wrong) on a play requires multiple viewings; ask your favorite NFL analyst if you don't believe us.

It gets even more confusing, though, when things like fantasy football are introduced. Now, don't get us wrong -- some purists hate fantasy football, but we're huge supporters. Anything that makes the Jacksonville-Cleveland tilt in Week 17 worth watching is good by us. But the problematic side of fantasy football is that it colors the way we view and understand the game.

In particular, fantasy football distorts the true value of running backs and running plays. In the majority of fantasy football leagues, a one-yard touchdown plunge is worth just as many points as a 69-yard run that gets bounced out of bounds at the one-foot line. Of course, that 69-yard run is far more difficult than the plunge; running backs pick up that one-yard TD on 55.9% of attempts.

That's an extreme example, but it points to the fetishization of touchdowns in most fantasy leagues, and that creeps into evaluations of how NFL players perform. Players like Marion Barber and Tim Hightower are said to have noses for the end zone because they accrue high touchdown totals, when the truth is that they really don't perform any better close to the goal line than an average back does.

Around the league this week, there are players who scored touchdowns, but contributed far more to their fantasy team's performance than the one that pays their bills. Matt Cassel threw for two scores, but they came down 24 points in the fourth quarter, and in the middle of a game where Cassel took five sacks and completed less than 50 percent of his passes while averaging four yards an attempt. Nate Washington caught seven passes for 66 yards and a score, but they came on 12 attempts (one of which was an interception), and included three empty completions that didn't help the team whatsoever.

The topper, though, was Kevin Smith. The Lions back scored twice on Sunday, converting from the 1-yard line and the 3-yard line. Valuable plays, sure, but not exactly hard work. He also had an 11-yard run. On his other 16 carries, he accrued a total of 15 yards. On the opening drive, his two carries went for -3 yards; the only reason he got to run the ball in from the 1-yard line to end the drive was because Chicago committed two penalties that gave the Lions a first down. All in all, eight of Smith's 19 carries went for no gain or negative yardage. That's an awful day, regardless of the work Smith did inside the five. Smith did something valuable by scoring on those two plays, but his poor performance on the team's other drives far outweighs the touchdowns, because scoring from inside the five isn't particularly hard to do. That's why he had -14 DYAR on Sunday, even though he earned his fantasy owners 17 points.

The lessons to learn: Don't mistake a good day in fantasy football for a good day in the real stuff. And don't confuse something that's important for something that's difficult.

Also, please note this week that adjustments for the quality of the opposition each player faces are now being factored into our numbers at 40 percent of their eventual weight. Each week, as we find out more about the true level of performance of each defense (and offense), we'll increase the weight of these opponent adjustments by 10 percent.

Quarterbacks
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
1.
Peyton Manning IND
31/41
353
2
1
198
198
0
Hey, we developed some advanced statistics to tell you that Peyton Manning was good. Can we have our medals now? Manning was 9-of-11 on third down, picking up six new sets of downs in the process. While some quarterbacks lock onto a specific receiver on third downs, Peyton kept 'em guessing: Three passes to Joseph Addai (including two dumpoffs), four to Dallas Clark, three to Reggie Wayne, and a bomb to Pierre Garcon that went for 35.
2.
Ben Roethlisberger PIT
26/33
333
2
0
194
190
4
Roethlisberger would have ranked ahead of Peyton Manning had it been he who threw the touchdown pass to Heath Miller, not Mewelde Moore. It's interesting that Mike Tomlin and Bruce Arians won't receive a single bit of criticism for their halfback pass inside the red zone, but the embattled Jim Zorn deserves to be fired for his. Good organizations -- both inside and outside of football -- judge performance based on process, not outcome. If you were calling for Zorn's head because of that playcall, Arians deserves your scorn, too.
MNF.
Brett Favre MIN
24/31
271
3
0
191
191
0
3.
David Garrard JAC
27/37
323
3
0
171
152
19
Garrard's 19 DYAR came on four carries: A 14-yard scramble on second-and-25, a 16-yard scramble on third-and-10, a successful sneak on third-and-1, and a 10-yard scramble on third-and-8. Meanwhile, Maurice Jones-Drew had six carries for 14 yards.
4.
Tom Brady NE
21/32
258
1
0
101
88
12
The most impressive drive of the season so far for Brady wasn't in Week 1; it was on Sunday, when he followed a sack that caused a fumble and a touchdown with this series: 4-of-4, 73 yards, and a 14-yard touchdown pass to Randy Moss on third-and-4. Speak of the devil, we were also amused on Sunday to see Moss lined up on the right, only to not run a route when the ball was snapped. Of course, he still drew double coverage, and the resulting space left a gap for Ben Watson to run a seam pattern for 34 yards through. Moss must've been loafing.
5.
Philip Rivers SD
21/36
254
3
0
85
89
-4
Our stats can adjust for the quality of a defense's play during the season, and the game situation that each quarterback faces, but not necessarily how an individual defense adapts to specific situations. And when the Steelers stopped getting pressure on Rivers and started dropping more defenders back into coverage, well, he picked them apart. Forget TMQ's "Stop Me Before I Blitz Again!"; we want "Prevent the Prevent!".
MNF.
Aaron Rodgers GB
26/37
384
2
1
67
62
6
6.
Drew Brees NO
20/32
190
0
0
65
60
5
Another victim of fantasy misperception; Brees didn't accrue some crazy level of production like he often does (thanks to Darrelle Revis just flat-out shutting down Marques Colston), but he still picked up nine first downs and -- most importantly -- didn't make any mistakes against the Jets D. No sacks, no fumbles, no interceptions. There's a mythology surrounding the "game manager" and what they do that usually revolves around simply having a great defense and not being particularly good, but what Brees did on Sunday is actually what a "game manager" is purported to do.
7.
Eli Manning NYG
20/33
292
3
1
59
59
0
We'll get to Steve Smith later, so with that in mind, Manning to receivers not named Steve Smith on Sunday: 9-of-18, 158 yards (97 coming on two completions), one touchdown, one interception, one fumble lost on a sack. Not awful or anything, but not exactly the stuff of legends against the Chiefs, either. Of course, David Carr was 0-for-2 and took a sack, so we expect that Giants fans will forgive Manning his transgressions if his heel would just heal up.
8.
Seneca Wallace SEA
33/45
257
1
0
47
39
8
Wallace actually had a pretty good game, with 16 first downs, a passing touchdown, and a seven-yard scramble for another one, but he was sacked five times and fumbled twice.
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
9.
Kyle Orton DEN
20/29
243
2
0
36
38
-2
Streaky streak. Orton started the game 2-of-8 for 13 yards with a sack, but once he threw his first touchdown pass, things started looking up. He had two different streaks of five straight completions, and he finished the game by going 6-of-7 for 97 yards, including the game-winning score to Brandon Marshall. Granted, he had little to do with Marshall's 33 yards after that catch, but we'll assume it was just Orton willing him to win in the huddle.
10.
Joe Flacco BAL
28/47
264
2
1
30
28
2
Flacco was great outside of one second-quarter stretch, in which he was sacked, picked up an intentional grounding penalty, had five passes hit the ground, and completed only one pass for five yards. If you see a highlight of Flacco's first touchdown pass, admire his ability under pressure; Flacco throws a perfect out to Derrick Mason for a touchdown despite having a rusher in plain sight about to lay him out.
11.
Matt Stafford DET
24/36
296
1
1
26
19
8
A week after we noted that Stafford enjoyed an incredible day on third down in the Lions' win over the Redskins, Stafford had a mixed bag of third downs before leaving with a knee injury. On one hand, he was 8-of-11 on third downs for 97 yards, picking up six first downs; on the other hand, he was sacked three times, fumbled the ball away once, and threw an interception.
12.
Shaun Hill SF
15/24
152
2
0
19
19
0
Winning 35-0 looks nice, but the 49ers defense scored three times; the offense only mustered 14 points against the lowly defense of the Rams, which isn't particularly impressive. Of note: The Rams sacked Hill four times after picking up only three sacks through the first three games of the year.
13.
Daunte Culpepper DET
6/11
54
0
0
18
18
0
14.
Carson Palmer CIN
23/44
230
2
1
16
10
6
Palmer threw 17 times on third down -- with an average of just 6.6 yards to go -- and could only pick up five first downs. Of course, two of those first downs came in overtime, with a third coming on a key fourth-down scramble, and that's when it mattered most.
15.
Derek Anderson CLE
26/47
269
1
1
14
5
10
Forget Braylon Edwards (no catches on five attempts); welcome to the big leagues, Mohamed Massaquoi! Anderson and Massaquoi attempted to hook up 13 times on the day, and while they completed six straight at one point, Massaquoi's last four targets fell incomplete, including a key drop. Correction: Welcome to the Browns, Mohamed Massaquoi.
16.
Tony Romo DAL
25/41
255
0
1
-3
-3
0
Never mind that Miles Austin was open on the game-tying touchdown. We're just confused. If you're the Dallas Cowboys, and you can spend over a billion dollars on your stadium, how can you end up throwing quick slants to Sam Hurd on the game's final two plays when he's being covered by Champ Bailey?
17.
Matt Schaub HOU
11/22
224
1
1
-5
4
-9
Nnamdi Asomugha did yeoman's work on Andre Johnson, but when Johnson was matched up against the unrelated Chris Johnson, the latter just sort of forgot he existed on Johnson's 62-yard pass play. Otherwise, Andre Johnson caught one of the seven passes thrown to him, picking up four yards on a third-and-6. That Asomugha isn't bad.
18.
Kerry Collins TEN
29/48
284
1
2
-11
-25
15
An exercise in attributing value to a quarterback: Take every story written about Kerry Collins last year. Focus on the ones where he was managing games, not making mistakes, and knowing when to step up his game on third down and the red zone. Note how he calmed the team down after the desperately poor stewardship of Vince Young... Now, read the stories about the Titans going 0-4. Are you going to read that the team is too hyper, or that Collins isn't managing games like he used to? Probably not. You're going to read that the Titans miss Albert Haynesworth and that their defense isn't playing the way it used to. Maybe that's a sign that the important variable last year wasn't Kerry Collins at all.
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
19.
Josh Johnson TB
13/22
106
1
1
-16
-24
8
Apparently, Josh Johnson paid really close attention to Jeff Garcia last year. Seven scrambles, but only one first down? Under five passing yards per attempt? He might as well be Vice-Captain Checkdown.
20.
Jay Cutler CHI
18/28
141
2
0
-26
-14
-12
21.
Chad Henne MIA
14/22
115
1
0
-42
-45
4
Henne got the win in his first NFL start, but that level of performance against a team that was missing nearly half its Week 1 starters on defense isn't anything to write home about.
22.
Matt Cassel KC
15/32
127
2
0
-68
-59
-9
Matt Cassel through three games last year: 53-of-89 for 458 yards, with two touchdowns and 10 sacks. Through three games this year, Cassel is 48-of-72 for 449 yards with two touchdowns and 10 sacks. The difference? Last year, he threw one pick in those three games; this year, he's thrown five.

EDIT: Cassel actually has five touchdowns against two picks. My bad.

23.
JaMarcus Russell OAK
12/33
130
0
0
-86
-82
-4
It's hard to pity an NFL quarterback making first overall pick money, but that's the emotion we have when we watch Russell play. The guy is clearly overmatched, and while he has a big arm, he simply isn't an NFL quarterback. Certainly not on the Raiders, at least. The scary thing is that the options behind him aren't any better.
24.
Jason Campbell WAS
12/22
170
2
3
-86
-94
8
Skins wideout Malcolm Kelly pulled a rare statistical trick this week: Two targets, two interceptions.
25.
Kyle Boller STL
13/24
108
0
1
-95
-74
-21
It turns out that Kyle Boller-to-Daniel Fells won't be the passing combination we'll tell our kids about for decades to come. Oh well.
26.
Trent Edwards BUF
14/26
192
1
3
-147
-147
0
One of the three picks was a Hail Mary, but Edwards was still subpar. Converting one of the ten third downs he faced isn't good enough; he only completed three passes on the penultimate down, and two of them were to the secretly-returned Marshawn Lynch for negative yardage.
27.
Mark Sanchez NYJ
14/27
138
0
3
-173
-182
9
Mark Sanchez looked great over the first three weeks of the year, but he's still a rookie. Occasionally, rookies ... well, they look like rookies. Peyton Manning threw three picks three times in his first four games, and he ended up okay. There will be days like this.


Five most valuable running backs
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
Matt Forte CHI
121
1
19
0
67
58
9
If you believed that there was something wrong with Matt Forte through the first three weeks of the season, you also probably believe that Forte solved it with his excellent performance on Sunday. There's no real difference between the Forte of Weeks 1-3 and the one who played the Lions on Sunday; it's the quality of the opposition. Backs will always look better against the Lions than they do against the Steelers. It's just how the world works.
2.
Rashard Mendenhall PIT
165
2
26
0
62
49
13
Point in case: The Chargers' defense looks woeful against the run without Jamal Williams. The result are huge holes for Rashard Mendenhall to run through, and a "breakout game" in the process.
3.
Ray Rice BAL
103
0
49
0
59
35
23
Baltimore's offense relies on a lot of play-action and deception with regards to their intentions in the backfield. That's dependent upon the abilities of Rice, who's equally comfortable carrying the ball up the gut as he is catching the ball in the flat. He even splits out as a wide receiver at times. He's not great at any of these individual tasks, but his ability to do them all makes him a very valuable cog in the Ravens' attack.
4.
Ronnie Brown MIA
115
2
0
0
48
48
0
Brown's day would look even better if he'd been more consistent; 64 of his 115 yards came on two carries. In all fairness, both those big plays were followed by short-yardage touchdown runs, and he came inches away from a third.
5.
Correll Buckhalter DEN
37
0
55
0
40
19
21
Buckhalter looks very good, and the Cowboys don't have a great run defense this year, but the Broncos' offensive line deserves a lot of credit for how that offense is playing. The front five of the Giants got a lot of hype last year, but Denver's starters were arguably as good and are playing like the best line in football this season.


Least valuable running back
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
Larry Johnson KC
53
0
1
0
-52
-29
-23
Johnson was having a quietly effective day as a runner until the Chiefs' final two drives of
consequence: Three carries, -12 yards, including a stuff on the one-yard line. He had no such excuse as a receiver, although the poor DYAR owes more to his role as a dumpoff guy than anything else: Three catches on five attempts for one yard.


Five most valuable wide receivers and tight ends
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Antonio Gates SD
9
12
124
13.8
2
59
The Steelers allowed starting tight ends only 35.9 yards per game last year, and were seventh in the league with a -14.8% DVOA produced by those players. They will not be hitting those numbers this year, although it took some incredible catches by Gates to make his 124-yard night happen.
2.
Calvin Johnson DET
8
12
133
16.6
0
53
One of the weaknesses of the Tampa-2 is that it doesn't allow for the same level of double coverage that is normally afforded star receivers like Johnson. One of its strengths is that it doesn't allow many big plays; unfortunately, Johnson proved to be the exception to that rule, starting off the game with a 45-yard catch and grabbing three more passes for 20 yards or more.
3.
Steve Smith NYG
11
16
134
12.2
2
50
At what point does he become the real Steve Smith and the Panthers one becomes the other guy? This Steve Smith's played in one more game, but he has 34 catches, 411 yards, and four scores; the Carolina Steve Smith has 15 catches for 190 yards and no touchdowns.
4.
Marcedes Lewis JAC
4
5
76
19.0
1
47
Lewis is a maddening player, because he's agile in the open field and can put up the occasional game like this to make it seem like he'll be a viable threat in the passing game. Then he goes and drops ten passes, like he did last year. Sometimes, you just have to take the good with the bad.
5.
Hines Ward PIT
8
11
113
14.1
0
43
Each year, people (including Football Outsiders) keep predicting that Santonio Holmes is going to take over for Hines Ward as the Steelers' primary receiver. And each year, Ward has continued to put up steady numbers, highlighted by games like this one. Maybe he deserves the benefit of the doubt until he proves he can't go.


Least valuable wide receiver or tight end
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Louis Murphy OAK
3
10
34
11.3
0
-38
In all fairness, the only reason JaMarcus Russell was throwing to Murphy and not Darrius Heyward-Bey is because Heyward-Bey can't run a route without falling down.

(Reminder: Quick Reads appears on ESPN Insider on Monday, then gets republished on FO on Tuesdays, with added ratings for Monday Night Football.)

Posted by: Bill Barnwell on 06 Oct 2009

92 comments, Last at 09 Oct 2009, 9:50am by Samoan_Rob

Comments

1
by Kulko :: Tue, 10/06/2009 - 11:56am

Is there really no way to compensate for RBs being the target of 35 dumpoff passes and having -3127 DYAR justthat the QB can avoid the sack?

Like having a separate baseline for RBs or such?

18
by Vincent Verhei :: Tue, 10/06/2009 - 1:14pm

Players are compared to how other players at their position do in similar situations. If Larry Johnson catches a dumpoff on third-and-20, that play is compared to other running backs when they are targeted on third-and-20.

19
by Independent George :: Tue, 10/06/2009 - 1:18pm

Somewhere in the Pacific Northwest, Mike Holmgren's ears are burning.

28
by C (not verified) :: Tue, 10/06/2009 - 2:12pm

When you calculate DVOA, how do you factor the magnitude of the the results?

How would you guys "score" a quarterback completing a 12 yard throw on 3rd and 10, vs a qb completing a 20 yard throw on 3rd and 10, vs a QB who completes a 60 yard touchdown on 3rd and 10. I mean, how much "extra" credit do the guys get with the longer plays and how do you quantify that magnitude?

If a make believe RB gets all 4 yard carries when the league average is 3 yards, vs a guy another make believe RB who gets either an 8 yard carry or a 0 yard carry every time?

Do you factor WHEN the each play took place? 1st quarter vs 4th quarter ( maybe even garbage time)?

Do you factor in what the score to the game is and what yard line the teams are on? For example, if the blue team is beating the red team by 24 points in the 3rd quarter and starts their drive 1st and 10 from their own 20, do you compare what happens on 1st and 10 "on average" in the NFL or what happens on average between teams up 24 points in the 3rd quarter starting from their own 20? There is a difference between playing defensive offense, running the ball, protecting leads and trying to score points quicker ( also in the reverse with the 2 min drill)... I mean 3rd and 4 with 1:54 left in the game when your team is down 30 points is very different than 3rd and 4 with 9 minutes left in the first quarter. You'd think by being very specific and narrowing everything down to time/space on the field & situation that it would vastly reduce your sample size. Especially since you compare everything against the current year right?

You said you don't factor in weather, but do you factor in grass vs turf, outdoors vs dome?

I'm a long time poster and I can honestly say I don't 100% "get" DVOA or whatever you guys are calling it now a days. I don't know what percentage of your readership knows everything about it exactly, but I know you guys don't reveal everything. I'm not embarassed to admit that either, I let you guys do your thing but I'm trying to see why some players might be better/worse than how DVOA scores them.

I was worried Aaron Rodgers would get short changed, but I guess you didn't rank him bad considering the impressive QB week with ( Big Ben, Peyton, Eli etc.)

31
by Eddo :: Tue, 10/06/2009 - 2:25pm

1. "How would you guys "score" a quarterback completing a 12 yard throw on 3rd and 10, vs a qb completing a 20 yard throw on 3rd and 10, vs a QB who completes a 60 yard touchdown on 3rd and 10. I mean, how much "extra" credit do the guys get with the longer plays and how do you quantify that magnitude?"

Success is not binary. A 20-yard pass on third-and-ten would have more value than a 12-yard pass.

2. "If a make believe RB gets all 4 yard carries when the league average is 3 yards, vs a guy another make believe RB who gets either an 8 yard carry or a 0 yard carry every time?"

This comes up all the time. There was an article a few years ago that concluded that a consistent running back is better than a boom-and-bust running back if their yards per carry figure was equal.

3. "Do you factor WHEN the each play took place? 1st quarter vs 4th quarter ( maybe even garbage time)?"

Yes.

4. "Do you factor in what the score to the game is and what yard line the teams are on? ... You'd think by being very specific and narrowing everything down to time/space on the field & situation that it would vastly reduce your sample size."

Basically, all plays are bucketed. Getting extremely specific would indeed reduce the sample size too much, as you note.

5. "Especially since you compare everything against the current year right?"

Plays are actually compared against a multiple-year average, but I'm not sure which years, exactly. That's why you'll sometimes see a leaguewide offensive DVOA greater than zero in a given year.

6. "You said you don't factor in weather, but do you factor in grass vs turf, outdoors vs dome?"

Weather's been a potential new variable for a long time, according to Aaron. Special teams numbers are already adjusted for weather, but it's very broad. The categories are warm, cold, dome, and Denver/Mexico.

40
by C (not verified) :: Tue, 10/06/2009 - 2:59pm

Eddo, thanks for the reply.

Getting 20 or 60 yards on 3rd and 10 is better than 12, but how do they quantify that magnitude?

I also fully understand that a cloud of dust RB is better than a boom bust RB (I've always argued this), but does DVOA ( or a personal players DYAR) reflect that preference and how so? Is Brandon Jacobs or Christan Okoye going to end up better off than the guy with negative and not much on plays and then long TD's?

If plays are bucketed regardless of the score of the game, won't that favor say players on bad teams racking up garbage yards over players on good teams with leads more or less running the clock out, say running on 8/9 man fronts? Running a 2 min drill on a prevent defense is a lot easier playing QB under normal circumstances.

I wish wind was factored in. A QB throwing passes with say 30 MPH winds in Chicago in decemeber has a harder job than being indoors in St. Louis even against the exact same defense.

Oh and one more question, aren't some subjectivly deemed "fluke" plays not counted for DVOA? Like say a hail mary and the end of a half? The QB won't get credit for the 50 yards and the touchdown?

I guess I'm sort of wondering how things are compared anyway... If on 1st and 10 the average run play is 2 yards, the RB gains 3 yards (50% better than average), how does that compare against say a 4 yard run ( 100% above average) or say a 17 yard run...

That would be cool of every situation could be compared to every situation, from the 46 yard line with offenses trailing by a field goal with 10 minutues left in the 2nd half, teams average 6 yards per pass etc. etc. etc.

47
by Bowl Game Anomaly :: Tue, 10/06/2009 - 3:28pm

Relative value of plays: I don't know the exact numbers, but I think 20 yards on 3rd and 10 is worth a small amount more than 12 yards, and a 60 yard TD is worth a lot more (but not 5 times as much).

Score: The score is not completely ignored. I think the categories are ahead/behind by 0-7, ahead/behind by 7-14, ahead/behind by 14+ or something along those lines. Time left is also factored, so garbage time stats don't count as much (although your definition of garbage time may differ from theirs).

Wind: I think they want to include wind but haven't figured out how to incorporate it in a way that makes the model better.

Hail Mary plays: Here's the thing about Hail Mary plays: INTs don't count, but TDs do. The reason should be obvious: A Hail Mary INT doesn't harm you any more than an incomplete pass, but a Hail Mary TD does benefit you. Hail Mary plays are desperation plays. You have nothing left to lose if you fail, but you can still gain if you succeed.

79
by NY expat :: Wed, 10/07/2009 - 12:07am

On the windy conditions issue, you've raised the issue many times, and I haven't seen anyone disagree that it *seems* like it should make a difference, but I also haven't seen anyone show something significant. Maybe it's not, overall, or maybe "windiness" is just too vague a term. After all, it's different if you going against the wind, with the wind, dealing with winds from either side, dealing with random gusts, etc. The actual wind speed would matter, but it would usually vary somewhat, maybe even play to play.

Also, playcalling presumably takes into account the windiness, so you might throw shorter passes, but still be able to get first downs and YAC so that the numbers don't look as bad as you might expect.

80
by tuluse :: Wed, 10/07/2009 - 12:20am

Please watch the "wind bowl" at Soldier Field in form the 2005 season or Brett Favre's "Last game in Soldier Field" from 2007 and tell me wind can't be significant.

82
by C (not verified) :: Wed, 10/07/2009 - 8:55am

But just look at any online sportsbook....

In some cities expected wind volitility might even be expected ( say Chicago), but when there is a really windy day in say Cincinatti, and the Over/Under changes by a few points changing the expected total score "projection" by 25% or so...

There is a "score predictions" market and DVOA is looking at past results not future projections right?

If the predicted score is expected to drop by 10%, 20% etc., couldn't you look at line movement in these crazy windy games?

Look at the Patriots/Buffalo game last year where Bellicheck didn't even attempt a pass in what the first quarter? Maybe even the first half? That was a crazy game. The wind totally changed the strategy of the game to the nth degree.

83
by dbostedo :: Wed, 10/07/2009 - 9:15am

I think maybe you've highlighted what the problem may be...wind isn't predictive of anything in particular, just that it has to be dealt with. For instance, perhaps coaching/team strategy makes for unpredictable scoring and yards.

I know that Aaron has mentioned trying to incorporate more weather data into the DVOA model, but everything it actually makes the model less accurate. So there may just be a need to tease out what the actual effects are. Does scoring really go down? Does completion rate go down? Do rushing yards really go up? Is it even the same depending on the particular wind conditions?

Without those answers, actually modeling wind effects could be impossible.

89
by C (not verified) :: Wed, 10/07/2009 - 4:51pm

Oh I never said it would be easy, but ignoring wind isn't the right thing either.

Rushing yards going up or rushing attempts?

Can you also even look at "scoring" and conclude it goes down? Expected scoring certainly does in high wind games. A fair comparison would be to compare the scoring between the Bills/Pats week 17 game with high winds and the same game without...

38
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 2:56pm

You might want to read (or reread) Hidden Game of Football by Palmer/Carroll. The basis of the play normalization in VOA is based on success points in that book, although it's been tweaked since then. If memory serves, the original system was "1" for a play that succeeded (40/60/100% of yards on 1/2/3&4 down) "0" for a play that failed, and then an additional 1 for 10, 20, and 40 yards. The cap at 40 is basically because anything past 40 has little to do with the player and far more to do with the location on the field. Can't run an 80 yard play at the 50.

The tweaking is mainly interpolating and then looking at more specific situations and tweaking things. Essentially the cap at 40 is due to the "depth of the defense," so if you can identify plays where the defense plays tighter, then that cap can go in. And plays where the defense is deeper (say, a prevent) could have the cap go out.

An argument could be made this isn't perfect, because you don't know how deep a defense 'really was' on any play. But that doesn't matter all that much, because you don't care about how the defense actually lined up, you only care about how an *average* defense would line up. If the Lions defense decided to constantly keep their safeties within 5 yards of the line of scrimmage, thus making all plays greater than, say, 10 yards equally easy, you're not going to adjust things to justify that stupidity.

"but I know you guys don't reveal everything."

The only thing they don't really reveal are the exact "success markers", which are pretty unimportant - they're just tuned to maximize the correlation. You could reproduce those yourself if you felt like it, or something very close to it.

The "theoretical" (if you will) basis for the rough values of those numbers comes from the game itself. Turnovers, for instance, are around -4 "real" points if you assume yardage is equally easy to gain anywhere on the field. With a bit more work you get a better estimate, but it's still pretty close to that.

43
by C (not verified) :: Tue, 10/06/2009 - 3:12pm

OK well here is where I can trace some of my differences on QB's...

If QB C keeps throwing 2 yard checkdowns on 3rd and 10... sometimes his back picks up the remaining 9 yards, sometimes not... He picks up 2 yards and leaves the remaining 9 yards to "chance". If he had Jones-Drew or Marshall Faulk back there, his stats and chance of "success" will be much higher than if he has Reuben Droughns or Kevn Barlow back there. Sometimes his guys will pick up the yards and give him "success", sometimes NOT. He did "part" of the work, but much of the "success" or "failure" is left out of his hands. He's passing the torch half way.

If QB E keeps throwing say 12 yard passes on 3rd and 10 and completing them... There isn't that "maybe my guy will pick up the extra yards for a success, maybe not".

The exact same result come about two different ways... Both greater than 10 yard passses result in "success" and first downs, the QB throwing the dig takes the chance out of it, while the guy throwing sideways to his RB leaves the chance in there. The QB with pocket hercules throwing check downs will do better than the QB with Reuben Droughns throwing check downs.

I understand you aren't going to factor everything into a given play... there isn't even the sample size to compare it against, but I am biased against QB's that pass the torch halfway, while I'm greatly in favor of guys who put themselves into position without having to do "extra" ( like picking up 80% additional rac yards to make the play).

That's just me trying to quantify the method to my madness.

48
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 3:34pm

If QB C keeps throwing 2 yard checkdowns on 3rd and 10... sometimes his back picks up the remaining 9 yards, sometimes not... He picks up 2 yards and leaves the remaining 9 yards to "chance".

Uh, I think the assumption that a back's ability to pick up YAC is independent of the quarterback's quality is really dubious. The fact that screens can be poorly thrown really makes that assumption unlikely.

But in any case, yes, so? That means that sometimes the QB will succeed on a 3rd and 11, sometimes he won't, and so his value would be less than a QB who throws a 12-yard pass consistently.

If he's got a great running back, yeah, he's gonna succeed more than an average QB, and so he's going to look better. But a downfield-passing QB with a great WR will *also* succeed more than an average QB, so he's also going to look better. Same problem - you're always going to have to infer who's most responsible, the QB, the WR, or the RB.

You do that by looking at similar plays with different players. Peyton Manning loses Marvin Harrison, loses Anthony Gonzalez, and doesn't miss a beat, and so we say hey, it's probably that Peyton Manning guy moreso than his receivers.

but I am biased against QB's that pass the torch halfway, while I'm greatly in favor of guys who put themselves into position without having to do "extra"

I think that's probably your own bias. A good example here is McNabb, who's a very good downfield passer. His short passing, however, is almost abysmal - some of his dumpoffs to Westbrook or other RBs in the past have just been near-comical. If you pretend that a 2-yard pass that picks up no yardage after the catch is the same for a quarterback who throws a 2-yard pass that picks up 10 yards after the catch, that's going to completely miss that point.

I think it's just safer to examine a quarterback's performance based entirely on the result of the play. Small stuff (like 'he checked down and the RB picked up the rest') is almost certainly going to be heavily biased. There's a ton of small things that a QB can do to make YAC easier, and it's impossible to figure in all of them.

With small sample size you have to qualitatively adjust things, but those things tend to even out over time.

58
by C (not verified) :: Tue, 10/06/2009 - 4:21pm

But most defenses play zone (man is just too hard), and 11 men trying to make a web that covers the field ( granted not in the redzone) is very very difficult to do. Even if you do have "good" or even "nearly perfect" coverage... A perfect throw can beat that good coverage. Teams don't play man as much. It's not like you have 1 WR up against 1 CB and he just beats him...

and even if you had a CB that was of equal talent of the WR, the advantage goes to the WR because he knows where he is going and the CB is reacting to him...

Playing CB is the hardest position on defense...

I don't think you can just take the attitude that "sometimes the guy throwing the check down will succeed, sometimes the guy throwing the 12 yard out will succeed... so what".

Yes, a 2 yard pass with 9 rac yards could gain 11 yards and a first down on 3rd and 10, and so can an 11 yard out... Same result, but I believe the QB throwing the out should be compensated MORE as it's a tougher play...

Throwing 2 yards and asking your back to pick up 9 yards is mostly dependant on the back. Sure, you could throw the pass a little in front of him to lead him ( as opposed to behind him or at him etc.), but you are really asking HIM to make the play.

If you throw an 11 yard out, throw the ball where it needs to be, and ask the guy JUST to catch it... A professional WR should catch the ball... Then the play is a success... it's less dependant on other people.

In the 11 yard out route, it's really the QB making the play
In the 2 yard swing pass and 9 yard run, it's really the RB making the play.

"I think it's just safer to examine a quarterback's performance based entirely on the result of the play."

Safer maybe, but fair I don't think so. Throwing a perfect swing pass to MJD who jukes, bounces off of, stiff arms, and out runs defenders for 40 yards is a lot different than throwing a nice 40 yard fade route down the sidelines. In once instance MJD made the play, in the other instance the WR made the play. What happens if MJD blows out his knee and now you have Greg Jones running that swing pass? You mean to tell me that QB isn't as good? I mean an "average" swing pass to MJD might even be better than a "perfect" swing pass to Greg Jones.

I agree that it would be hard to rank how accurate a pass is and not even all completions are "accurate" ( throwing behind a guy instead of leading him).

Mcnabb is good because he's smart, consistant and makes the right reads... I don't think most people know he studies so hard and is a geek in the film room. Accurate throws aren't his strenght but I think his accuracy has improved ( remember those ground balls he used to throw at people's feet early in his career?)

I'd take a smart (consistant) QB that isn't super accurate like Mcnabb ( Giants #10) because knowing WHO to throw the ball too is the hard part. There are plenty of strong armed and even accurate QB's that make huge mistakes turnover producing mistakes. If you made just 2 stupid picks every game (but did pretty much everything else right), then you aren't a good quarterback. Think about that. You drop back 25 times and could theoretically do the right thing 92% of the time, but those 2 times you screw up... make you a bad player.

62
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 4:52pm

Yes, a 2 yard pass with 9 rac yards could gain 11 yards and a first down on 3rd and 10, and so can an 11 yard out... Same result, but I believe the QB throwing the out should be compensated MORE as it's a tougher play...

You believe it's a tougher play. I don't buy your argument. I don't see why delivering a screen with a lineman in your face to a receiver that's *actually* about 10 yards away (although 2 yards downfield) is "easier" than throwing a 10 yard pass when the receiver is stationary and directly 10 yards in front of you with a clean throwing lane. That's an exaggeration, yes, but the point's still the same.

I don't think you can blindly classify "short" throws as easy and "long" throws as hard. Especially when you're classifying "short" and "long" as distance from scrimmage.

Again, take a screen play - if the screen's set up well, all the running back really has to do is follow his blocking. A professional running back should be able to do that. The difficult part is getting the timing right, and that's the QB.

Playing CB is the hardest position on defense...

Don't believe that either. A given corner is only involved in a play on maybe half the snaps he sees. I'd say linebacker and safety is probably harder to play because you have to diagnose the play more. Corners, especially playing man, really only have to focus on their guy and their responsibility.

If you throw an 11 yard out, throw the ball where it needs to be, and ask the guy JUST to catch it...

No. That is absolutely, 100%, not all that the WR is asked to do. The WR had to run the route in the first place, and that is what determines whether or not the play succeeds. If on 4th down, the WR runs his route a foot inside the first down marker, the play's a failure even though the QB did everything right.

I think assigning responsibility purely to the QB in the case of a 10-ish yard pass and YAC purely to the RB in the case of a screen pass just doesn't stand up.

In once instance MJD made the play, in the other instance the WR made the play. What happens if MJD blows out his knee and now you have Greg Jones running that swing pass?

The same thing that happens if a star WR blows out his knee and is replaced by Billy McMullen. In the first case, screens/swings/etc. become less effective. In the second case, 12-yard outs become less effective.

but I think his accuracy has improved ( remember those ground balls he used to throw at people's feet early in his career?)

Yes. I saw it Week 1 in 2009. His accuracy hasn't significantly changed. The pass distribution and quality of his receivers has. He still throws low all the time.

I'd take a smart (consistant) QB that isn't super accurate like Mcnabb ( Giants #10) because knowing WHO to throw the ball too is the hard part.

This is just boom/bust vs. consistent production all over again. Both can be successful, but only if the risks that the boom/bust QB are successfully mitigated by the gains that he gets.

74
by C (not verified) :: Tue, 10/06/2009 - 6:42pm

I wasn't talking about "just" a screen pass. If the QB pulls a Jaguars Byron Leftwich and drops back and throws sideways to his RB... Maybe literally sideways, maybe a 1 yard throw, maybe a 2 yard throw... more of a check down... You could disagree with me or not, but that SHOULD be a very easy throw for even an average quarterback. I know you'll argue with me, but any decent quarterback should be able to throw sideways + or - a couple yards to his back. Yes, screens take timing, but throwing a check down should be a very routine play. After you get the ball to your back... they are the one making the play after that... The QB shouldn't have had to make a tough "read" or anything like that. It's a brainless play and execution.

You are right, it isn't a WR JUST making a catch, he has to run the route, but much of the risk of throwing that 12 yard out, dig, come back or whatever is the QB reading the defense... making sure the guy is in between zones or open. That's the hard and potentially dangerous part of the play. There is more risk but more reward... These passes are much more likely to be intercepted than FB check downs. There is more risk however if you have a good QB ( Peyton Manning) his intelligence can mitigate this risk.

The Bill Walsh school of thought says you can use smarts on offense, but on defense you need the physical athletes.

You admitted yourself that Peyton Manning could say "add value" to his WR core. So why doesn't he get credit for adding that value? Why does the guy throwing check downs get just as much credit as the guy taking more risk ( for more reward), but without getting burned on that risk he's taking?

If throwing the ball down field was just as easy as "short throws". Then how come young players can't do it as well without turning the ball over? Matt Stafford clearly has the arm to throw throw 12 yard outs and digs...

Further a QB throwing the ball only down field would be harder to defend than a QB who only throw the ball within 5 yards of the LOS.

Corners playing man "only" have to focus on their man true, but covering somebody even just as fast as you is very very hard... Covering somebody faster/taller/more agile than you is even harder. There is a reason why GM's say there aren't anymore man to man corners in the NFL... because it's super a difficult assignment.

76
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 8:27pm

After you get the ball to your back... they are the one making the play after that...

Except the QB set the play up for them. It's exactly the same as giving all the credit to a RB for the run and ignoring the offensive line.

Even a checkdown can be set up by the QB, if the QB moves safeties, or if the QB buys time, or heck, if the QB draws defenders away from the back.

The QB shouldn't have had to make a tough "read" or anything like that. It's a brainless play and execution.

1) Wait, a checkdown *is* a good read. By definition. A checkdown happens when there's nothing else available and the QB has let the play develop. It's not brainless. Brainless play would be Rex Grossman saying "F* IT, I'M GOING DEEP" even though Berrian is triple-covered downfield instead of checking down to a wide open RB.

2) By that definition, there are brain dead 10-yard outs, too.

I think the problem here is that you've got an image of a QB checking down to a receiver who's wide open 2 yards away from the LOS, and then somehow going all Barry Sanders and picking up a first down by weaving through traffic. Is that right?

In that case, the QB would be getting credit for what's primarily the receiver's work, yes. So is that a problem? No, not really, because an *average* receiver isn't going to pick up the first down that often, and heck, even some super-RB isn't going to pick up the first down all that often either. Which means yeah, that play *is going to work out to be worse* than a route right at the first down line.

I mean, really, if you're checking down to a FB, how often is that back going to pick up the first down? Not freaking often. And if he does, and the QB gets credit? Big deal. It's exactly the same as the QB getting credit for a receiver who makes a fantastic grab on a terrible throw. It happens. Sometimes it's obvious. Sometimes it's not.

Yeah, the QB looks better with a more agile receiver. Why is this a big deal? Daunte Culpepper looked like a super-QB when all he had to do was chuck it near Randy Moss. *We know* that good receivers make QBs look better. You *always* have to deal with that problem.

You admitted yourself that Peyton Manning could say "add value" to his WR core. So why doesn't he get credit for adding that value?

Because the only way you can determine that it's Peyton Manning and not the WR is by noting that the performance doesn't change when WRs change. You don't have that luxury on individual plays. And even then, you'd prefer to see the WRs go to another team and see their performance drop.

If throwing the ball down field was just as easy as "short throws". Then how come young players can't do it as well without turning the ball over?

For the same reason young players throw pick-6s on quick slants. Because they're young and make mistakes.

There is a reason why GM's say there aren't anymore man to man corners in the NFL... because it's super a difficult assignment.

I think we're just disagreeing as to what "difficult" means. I think corner is the position that requires the most speed and agility on defense, yes.

81
by C (not verified) :: Wed, 10/07/2009 - 8:50am

"1) Wait, a checkdown *is* a good read. By definition. A checkdown happens when there's nothing else available and the QB has let the play develop."

But how do you really know nobody else was open? The TV doesn't even show much at all past the lines. I don't think you can assume that "nobody downfield" was open, that's a fallacy people fall into.

I mean, look at a Brett Favre or Jay Cutler... You even get to the point where defenders could be draped all over their receiver but they are so confident in their arms and abilities (risk tolerence) to fit the ball in tight spaces... Open to them might be very different to a risk averse quarterback, or a quarterback who is specifically told by his OC or HC NOT to take chances.

If an average WR runs the 12 yard out (correctly)... then as a professional athlete he should make the catch... Assuming it isn't some crazy sideline catch etc. Then QB did the "hard part" of the play by identifying the coverage and making the throw. Yes I know the O-line had to protect him and such, and I know he might have a bad throw... but I believe if you get your hands on the ball as a WR, you make the catch.

Throwing a 0 yard swing pass is a much easier play. Now if you want to argue the percentages of picking up the yards... that's not what I am arguing... I'm saying it's an easier play to make, and therefore shouldn't receive the same credit as throwing for the yards through the air. That's the harder play to make, that's the harder play to defend, and that should receive more credit.

84
by dbostedo :: Wed, 10/07/2009 - 9:28am

C - I think maybe you're getting caught up in thinking about single play value, when the system here is designed to assign value to every play, in every situation. That means that there simply isn't room to tease out details and think about particular situations that might give weird results.

Yes, those situations exist and maybe value isn't attributed correctly. But things should even out over larger samples, like any statistical model.

And yes, all of the DVOA values are somewhat dependent on teammates. Is it the running back or the o-line? Is it the QB or WR? Is it the TE or crappy LB coverage? You might be able to look at individual plays and glean those answers. But if you're going to come up with a system to grade players, there just no way to statistically separate that stuff out. That's one of the reasons this is all so hard...

90
by C (not verified) :: Wed, 10/07/2009 - 5:02pm

That's a single play value, but what if one player has a trend on said particular behavior.

Let's say the Jaguars have a quarterback who does two things (exaggeration). He throws a swing pass most of the time and maybe twice per game he throws a predetermined deep ball... Then a few years later he's a journey man bouncing around different teams and he doesn't get nearly the positive reviews he used to...

Sure, you can argue that on any given swing pass he might be given extra value, but what if the guy keeps doing that over and over again? What if his game IS the swing pass?

I'm trying to think of an analogy. It would be a like having two bench baseball players with limited at bats. One who kept getting hits off of other teams #1 starters vs a guy who kept getting hits off of garbage middle relievers. If you were just going to look at batting average that's one thing, but look at how much more impressive it is to get hits off CC Sabathia instead of Kyle Farnsworth... I know I know, DVOA factors against competition but I think it's off in that case.

Yes people do grade players, it is subjective, and some people are good at it... The whole NFL world doesn't revolve around statistical analysis. Some front office guys do have an eye for talent ( or trusted scouts)... Newsome, Reese, Smith, Butler, Polian, Pioli..

91
by Sifter :: Thu, 10/08/2009 - 1:16am

"Sure, you can argue that on any given swing pass he might be given extra value, but what if the guy keeps doing that over and over again? What if his game IS the swing pass?"

Well the defense would sell out to stop the swing pass and it would soon become a very inefficient pass for Mr Campbell..err I mean Mr Hypothetical short passing QB...

Think of New England. Brady to Welker is an insanely successful SHORT passing combination and every time I watch a Patriots game it drives me insane how easily Welker gets open and gets first downs. But it's because of the deep threat. If there was no Moss or Brady had never thrown a successful deep pass, Welker would be swamped. But when Brady throws to Welker it is almost the DEFENSE CHOOSING to leave open the lesser of two evils. That is why I mentioned an adjusting defense straight away, a short passing QB will soon find his zones shortened and safeties playing close to the line. He won't be able to keep it up unless he has a deep threat.

So back to Jason Campbell, the fact that he's completed a couple of long passes to Santana this year at least makes the defense back off a little bit. So in that way he deserves at least a little credit for making the defense think twice and hence keeping those underneath checkdowns open. So to me, as long as a QB is actually MOVING his team and getting first downs, I don't care what degree of difficulty his passes have. It's all about efficiency not style...

78
by Jerry :: Tue, 10/06/2009 - 10:42pm

The play-by-play just says "Smith pass short right to Jones for 12 yards"; it doesn't specify how much is YAC. So DVOA would have problems trying to include that.

46
by dmb :: Tue, 10/06/2009 - 3:21pm

Chris, the answers to most of these questions are readily available on this very website.

"[DVOA] uses a value based on both total yards and yards towards a first down, based on work done by Pete Palmer, Bob Carroll, and John Thorn in their seminal book, The Hidden Game of Football. On first down, a play is considered a success if it gains 45 percent of needed yards; on second down, a play needs to gain 60 percent of needed yards; on third or fourth down, only gaining a new first down is considered success.

"We then expand upon that basic idea with a more complicated system of 'success points.' A successful play is worth one point, an unsuccessful play zero points. Extra points are awarded for big plays, gradually increasing to three points for 10 yards, four points for 20 yards, and five points for 40 yards or more. There are fractional points in between. (For example, eight yards on third-and-10 is worth 0.63 "success points.") Losing four yards is -1 point, while losing 12 yards is -1.8 points. Interceptions average -6 points, with an adjustment for the length of the pass and the location of the interception (since an interception tipped at the line is more likely to produce a long return than an interception on a 40-yard pass). A fumble is worth anywhere from -1.70 to -3.98 points depending on how often a fumble in that situation is lost to the defense -- no matter who actually recovers the fumble. Red zone plays are worth 25 percent more for teams (and 10 percent more for players), and there is a bonus given for a touchdown.

"Every single play run in the NFL gets a 'success value' based on this system, and then that number gets compared to the average success values of plays in similar situations for all players, adjusted for a number of variables. These include down and distance, field location, time remaining in game, and current scoring lead or deficit. Teams are always compared to one standard, as the team made its own choice whether to pass or rush. However, when it comes to individual players, rushing plays are compared to other rushing plays, passing plays to other passing plays, tight ends get compared to tight ends and wideouts to wideouts..."

If you want to read more about it, there's more than a dozen paragraphs explaining DVOA right here: http://www.footballoutsiders.com/info/methods#DVOA. (Or you can simply put your cursor over the "Statistics" tab, and click on "Our New Stats Explained." It's the first option under statistics, and it's even highlighted!

2
by Independent George :: Tue, 10/06/2009 - 11:57am

If I'm not mistaken, Eli's numbers passing to other receivers includes two drops by Manningham (one of which resulted in the interception). 11-of-18 for, say, 170 yards, 1TD, 0 INTs seems pretty efficient to me.

3
by andrew :: Tue, 10/06/2009 - 11:59am

I'll wager Adrian Peterson was near the bottom of the league this week.... that was his worst game since his rookie game against the Niners. He had over 20 carries, and I remember him converting one 3rd & 1, one TD from first and goal at the 1... and a first down on a 12 yard run. The rest were mostly failures.

Was it the packers overplaying the run (and thus the Favre 190 pt game) or is it something else...

6
by Will Allen (not verified) :: Tue, 10/06/2009 - 12:17pm

Peterson seemed to start o.k., fumble excepted, and then the Packers seemed to decide that they were going to fill all gaps, no matter the cost in coverage or pass rush, and Favre was quite happy to carve them up as a consequence. I can't go back and check, since a non football fanatic in my house deleted the recording (!). It'll be interesting to see how defensive coordinators react if Favre, Berrian, Harvin, Rice, etc., continue to have substantial success downfield.

9
by Bobman :: Tue, 10/06/2009 - 12:26pm

Will, Sorry you don't have it recorded--that would be a nice game to hold on to. While Favre looked good as you point out, the D was stunning. Awesome goal line stand.

That's quite an... intimidating team right now. A few teams are clicking on all cylinders as of Week 4--I wonder if they'll all be able to keep it up.

60
by Tundrapaddy (not verified) :: Tue, 10/06/2009 - 4:43pm

I missed it entirely. I alternated between Gamecast and AM radio.

It certainly appeared (from the non-visual info I could gather) that the Packers were selling out to stop the run, much like the Titans did against J-ville.

And the end result was the same (MNPQ and Garrard shredding the Green Bay and Tennersee secondaries).

The downside is that I had both Peterson and 'Pocket Hercules' (MJD) as my running backs in my fantasy league.

Anyhow - I was thrilled to hear the goal-line stand, although it seems that the Minny defense is slipping, overall, from what they were last season. Is age finally affecting Pat Williams' game?

61
by M :: Tue, 10/06/2009 - 4:51pm

Peterson & Pocket Hercules? BOTH?!?!?!? I have no sympathy for you whatsoever; my intended top 2 RB's are Jacobs & Tomlinson. I am desperately hoping that both teams start using these backs for goal-line carries; frankly, both of them seem to be missing something this year. I just hope that the schedule does get a bit easier for both of them down the stretch.

Regardless, even if they both "progress" to the mean, I chose badly on draft day.

65
by jds (not verified) :: Tue, 10/06/2009 - 5:03pm

Is age finally affecting Pat Williams' game?

I'm also worried about Kevin Williams. Where has he been the past few weeks?

75
by Arkaein :: Tue, 10/06/2009 - 7:18pm

With regards to last night's game, I think GB and Minny had exactly the opposite defensive game plans for their front sevens. GB sold out to fill every running lane and shut down AP, which they did but to the detriment of their pass rush. Minny on the other hand sold out to attack Rodgers and dared GB to beat them with the run.

GB's running game was fairly effective, just not enough to compensate for the tissue paper pass protection. I would take Minnesota's lapse in run defense for that game at least with a grain of salt, for the same reason I don't think that GB's pass rush is really as bad as it looked last night. Defensive schemes had a lot to do with both.

63
by M :: Tue, 10/06/2009 - 4:56pm

I listened to the game while working (very) late; if the Packers had even a Division I caliber offensive line, does it seem like they could have stolen the game at the end? From what I was hearing, there was a lot of running AP into a pile of people and disinterested coverage of Packer receivers by the secondary. If anyone can shed some light on this, I'd be appreciative.

I'm hoping that we don't have such a half-assed "run out the clock" strategy if we're playing a team like the Giants, Eagles, or Saints. I think we would quickly lose the lead playing "Chilly-ball".

72
by drobviousso :: Tue, 10/06/2009 - 5:45pm

"if the Packers had even a Division I caliber offensive line, does it seem like they could have stolen the game at the end?"
Yes and no. Yes because if they had any protection, they probably won't have needed two last minute scores. No because they didn't recover either of the on sides kicks, and that was more about Sidney Rice making 2 plays than the O line.

4
by andrew :: Tue, 10/06/2009 - 11:59am

weird. somehow I double posted an extra blank message. I edited it to put this text in it. Please disregard.

5
by BroncosGuy (not verified) :: Tue, 10/06/2009 - 12:09pm

Bill, you make two good points here.

The first, I think, most people realize: fantasy points are distorted by role. You make a lot more of those 1-yard TD plunges if that is your role. And if that is your role, it does not make you a "better" back than the guy who got you to the 1-yard line.

The second stems from your Larry Johnson comment (I didn't see the Chiefs game so I'm relying a bit on inference, but the point should still hold). FO stats are also clouded by usage or, more accurately, the talent and decision-making of teammates and coaches. If a QB checks down on 3rd-and-23 to a 4-yard swing pass, FO views this as a rather useless play. It does not mean the receiver "sucks" or that he is less talented than a player with a higher DVOA or DYAR. Aaron inserts this disclaimer (far more eloquently) in every book, but I still see a lot of comments suggesting DVOA proves player A is better than player B. It does no such thing. Just something to keep in mind.

36
by countertorque :: Tue, 10/06/2009 - 2:49pm

DYAR is dependent on the players role and the number of touches he gets. A low DYAR does not by itself indicate that a player sucks.

DVOA calculates how well a player does compared to his peers in similar situations per play. It is not a perfect relative measure, as it cannot account for everything (eg. quality of teammates). It is the best thing we have. If a player has a very low DVOA, it means he's done a lot less per play than his peers in similar situations. That's a pretty good measure of sucking, IMO.

7
by t.d. :: Tue, 10/06/2009 - 12:17pm

Eli is solidifying his position among the upper echelon of quarterbacks

8
by Bobman :: Tue, 10/06/2009 - 12:22pm

It was my distinct impression that Seneca Wallace's yardage and scores were in garbage time. There is a correction for this, no? Though, I guess defining garbage time is difficult--28 pt deficit? 24 pts? Dunno.

Does Aaron Rodgers get some kind of credit in his account for the JV OL he had? Not sure if P Manning caught a similar break two years ago against SD when half his OL and WR corps went down (resulting in 6 INTs), but Rodgers was just toast all night--to the point that, down 14 late in the game, they ran 3-4-5 straight times to the right side because the left side of the OL was so iffy. Instead of bombs away inside their own 5, they elected just to give their punter breathing room. It proved to be a reasonable strategy as they had chances to tie/win in the end.

On the flipside, Wallace's troubles came largely from a JV OL, but he kept trying to pass.

11
by Led :: Tue, 10/06/2009 - 12:33pm

GB's pass blocking was poor, but Rodgers held the ball way too long on several of those sacks.

56
by J.D. (not verified) :: Tue, 10/06/2009 - 4:11pm

Have Rodgers and Rob Johnson ever been seen in the same room at the same time?

49
by coltrane23 :: Tue, 10/06/2009 - 3:41pm

Yeah, I was surprised to see Seneca Wallace so high on the list. I guess that my unscientific impressions from the game speak to Manning's effectiveness, because when I compare Wallace's game to Manning's game in my mind, Wallace's game seemed like the equivalent of empty calories.

Sadly, I knew that game was over after the Colts' first TD drive.

53
by Bobman :: Tue, 10/06/2009 - 3:56pm

For me it was the 78 second drive just before the half. I was pretty sure it would happen, but a stumble, or even a FG might have left the door open. But for most of the game, the TV shots of the Seahawks bench were just depressing--those guys looked like their puppies were just eaten by the neighbor's cat. NOBODY looked like they wanted to go back out on the field.

10
by Jimmy :: Tue, 10/06/2009 - 12:28pm

The Bears actually did a much better job of controlling Megatron in the second half when they used much more tampa 2 coverage. What killed the Bears in the first half was leaving Bowman on an island against him when they were blitzing. He is always going to be a threat on the quick fade behind the corner and the slant in front of the CB but at least you are keeping a safety behind him most of the time. In the second half the Bears used Tillman to battle him off the line in the cover2 shell and held him to 14 yards (I think and that might be missing a penalty). That will only work if you have a corner who is able to play as physically as Tillman though.

12
by JasonG (not verified) :: Tue, 10/06/2009 - 12:40pm

Two sacks and another fumble, but also a TD run on 3rd down (no extra credit for style points, I know). That's a negative Rush DYAR?

His passing numbers weren't gaudy, but still a high percentage with 2 TDs and no INTs, which came to a 100.4 rating for the game. The only negative I can think of is poor third down conversions. That's equates to a negative Pass DYAR?

High percentage, three TDs, one fumble adds up to a below average performance?

14
by Eddo :: Tue, 10/06/2009 - 12:55pm

1. He averaged barely five yards per pass attempt. That's pretty bad.

2. I'm assuming the second fumble, and subsequent loss of yardage, counts against Cutler's rushing DYAR (if not, it should, that didn't look like a pass play to me). I also remember a second-down scramble that only got about 50% of the required yardage for the first down. That will hurt his rushing DYAR more than the touchdown, impressive as it was, helps it.

In regards to that second fumble: Cutler's ankle was stepped on by Kreutz before he could even take his first step. As my brother said while we were watching the game, at what point does the Bears' staff realize that aborted snaps and quarterbacks getting their feet stepped on has been a chronic problem for years, regardless of who's taking the snap? Kreutz is having a good year overall, but he's always had issues with this sort of thing.

Overall, Cutler's DYAR matches my impression of the game; he didn't look good, but didn't really have to, but he also didn't look too bad. A very "meh" performance from him, I thought, especially after two impressive weeks.

15
by Jimmy :: Tue, 10/06/2009 - 12:57pm

A very "meh" performance from him, I thought, especially after two impressive weeks.

Isn't it fun to have a QB who's 'meh' performances don't involve four turnovers and at least one twenty yard sack.

35
by Charlie (not verified) :: Tue, 10/06/2009 - 2:43pm

It so is.

39
by TomC :: Tue, 10/06/2009 - 2:58pm

Amen, brother.

22
by Vague (not verified) :: Tue, 10/06/2009 - 1:27pm

Im on the replace kruetz bandwagon.

68
by Jimmy :: Tue, 10/06/2009 - 5:11pm

Kreutz is (stepping on his QB's foot once aside) playing well this year. Unlike some of the other linemen he seems assignment sure and is still a top center. There has been - to my recollection - only one other snap snafu so far (botched snap in the first game) and that stuff can happen, although I would concede that he isn't the most perfect of snappers.

71
by Thanos (not verified) :: Tue, 10/06/2009 - 5:37pm

I'm on the get a new OL Coach bandwagon!

13
by Jimmy :: Tue, 10/06/2009 - 12:54pm

Cutler actually fumbled twice. The first wasn't really his fault ,the blitz pick up simply has to be better than it was on the strip sack. The second was when Kreutz trod on his foot as he came out from under center; he fumbled as he staggered sideways although he didn't fumble very far and was able to fall down on the ball. It was still a fumble though and until Cutler stops doing it he isn't going to generate a lot of DYAR and rightly so to be honest. Both those balls could easily have gone to the opposition and had a negative impact on the game.

That and the Bears were crap on third down.

Does anyone know what is wrong with Olsen? Yet another game with loads of drops or incompletes.

41
by tuluse :: Tue, 10/06/2009 - 3:03pm

I've noticed that Cutler has poor ball security. and a touch too much faith in his protection. However, the latter is probably a good thing. Otherwise you can end up looking like Tony Romo did against the Broncos.

The offensive line has to be the point of emphasis this offseason. We have a QB worth protecting, our receivers have proven to be good enough, there are a couple defensive needs, but getting Cutler alive and making sure he doesn't get jumpy need to be the biggest concerns.

70
by Jimmy :: Tue, 10/06/2009 - 5:32pm

I agree with you about the offensive line it would be great to see Cutler if he had a load of time and a strong running game. However I fully expect that we will see exactly the same group of linemen as this year next year. Pace and Williams with Schaffer as the swing tackle are going to be the tackles, Kreutz will be the center and Garza and Omiyale will play guard. They will keep Beekman around and liked Louis enough to keep him on the roster not the practice squad which was a bit of an eye opener. With the team liking Louis and probably wanting to hold onto Beekman they will probably only take a lineman on the first day if a guy they love slips. If Pace retired they wouldn't draft a tackle they would draft a guard and move Omiyale to RT. I am not saying I love the idea but I don't expect the Bears to bring a whole load of offensive linemen in.

They'll end up taking a DE and a CB with their first two picks.

44
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 3:14pm

"he fumbled as he staggered sideways although he didn't fumble very far and was able to fall down on the ball."

That's really luck, though. Ball bounces wrong, and it's out of his reach and there's nothing he can do to recover it because he's off-balance.

I *hate* seeing OL step on a quarterback's feet. Not only does it have very bad injury risk (imagine especially if the offensive linemen gets off-balance and gets bull-rushed by the defensive lineman, falling on the quarterback's leg) it kills the mobility of the person who's actually most able to recover the ball if he loses it.

66
by Jimmy :: Tue, 10/06/2009 - 5:07pm

I agree with you, fumbling is bad. He didn't fumble as he took the snap, he just fell down when Kreutz stepped on his foot. The dumb mistake was that he tried to get up and run around left end fumbling in the process.

16
by Joseph :: Tue, 10/06/2009 - 12:59pm

JasonG,
Remember, Aaron now is factoring in DEFENSES. Since Cutler's day came against the LIONS, my guess is the D in DYAR is contributing to most of that negative rating.

24
by JasonG (not verified) :: Tue, 10/06/2009 - 1:36pm

I didn't realize he had been charged a fumble when his foot was stepped on and he tripped. I thought it was just a lost yardage play. Also, it's Week 4, so I'm beyond reading the intro thoroughly and missed the opponent adjustment. That certainly makes a difference when facing a bottom 5 pass defense.

17
by Israel P. (not verified) :: Tue, 10/06/2009 - 1:04pm

When you start adjusting for defense, how does the fact that Troy didn't play affect Gates' performance? Whatever the Steelers will be then, they clerly cannot be that now. And Gates is Troy's guy.

20
by TGT2 (not verified) :: Tue, 10/06/2009 - 1:23pm

DYAR does not adjust for Injuries in the same way it doesn't adjust for scheme.

27
by Israel P. (not verified) :: Tue, 10/06/2009 - 2:04pm

That was rhetorical.

21
by bravehoptoad :: Tue, 10/06/2009 - 1:26pm

The 49ers defense only scored 2 touchdowns; 1 was scored by special teams.

Together with 2 on offense, this is why people are calling it a "complete game" (if they're being optimistic.)

26
by Independent George :: Tue, 10/06/2009 - 2:02pm

The game appears to have gone the full 60 minutes; surely that qualifies as a complete game.

54
by Bobman :: Tue, 10/06/2009 - 3:58pm

As my father would say, very funny, wiseass.

73
by drobviousso :: Tue, 10/06/2009 - 5:52pm
23
by chrisV (not verified) :: Tue, 10/06/2009 - 1:29pm

You got Cassel's stats backwards, he has 5 touchdowns and 2 interceptions.

25
by Bill Barnwell :: Tue, 10/06/2009 - 1:48pm

Oops. You're 100 percent right. Sorry about that, will fix.

29
by Rich Conley (not verified) :: Tue, 10/06/2009 - 2:14pm

"Baltimore's offense relies on a lot of play-action and deception with regards to their intentions in the backfield. That's dependent upon the abilities of Rice, who's equally comfortable carrying the ball up the gut as he is catching the ball in the flat. He even splits out as a wide receiver at times. He's not great at any of these individual tasks, but his ability to do them all makes him a very valuable cog in the Ravens' attack. "

How much of Rice's DYAR on the day was the play where Guyton blew the angle and he went for 50+ yards?

I'm in Chicago now, and Kreutz really needs to go. He's not a very good blocker anymore (he gets pushed back more often than not), and he can't shotgun snap.

30
by Samoan_Rob :: Tue, 10/06/2009 - 2:19pm

Typical fantasy league scoring:

1 point for every 10 yards rush/rec
6 points for every touchdown
0 points for every first down

I know any answer is far from DVOA, but let's start here:

What integers x, y, z best correlate with DVOA:

x points for every 10 yards rush/rec
y points for every touchdown
z points for a first down

Baseball boxscores ignore walks and are stuck w/wins, losses.
Let's evolve fantasy scoring before it's too late.

33
by Eddo :: Tue, 10/06/2009 - 2:28pm

P-F-R attempted to do something like that in their ranking-QBs-by-peak post last week (which was linked in an Extra Point here).

77
by Samoan_Rob :: Tue, 10/06/2009 - 10:41pm

Thanks. Guess I was getting at the fact that I've never seen a fantasy league that rewards first downs. Yet, it's in the CBS Sportsline setup. You could do it. I just wonder if 1 pt per 10 yds, 2 pts for a first down and 6 pts for a TD is 'right' or if it should be 0.5, 3, 3 or something.

85
by dbostedo :: Wed, 10/07/2009 - 9:39am

I doubt 6 points is the best value. I think one of the more obvious disconnects between fantasy performance and "real" value, however you define it, is touchdowns by running backs who primarily get the ball near the goal line.

I've been arguing with my league members that TDs should simply be a bonus - maybe a point or two - and that yards should be all important. I suppose adding in first downs would potentially be even better, but I don't know for sure.

I think the secondary problem is that a league based on real value might not be as much fun. You might wind up with situations where, for instance, a good day by a QB could almost outscore an entire other team. So to some extent, you might not want your league to reflect too much reality.

86
by M :: Wed, 10/07/2009 - 10:56am

Hear ye, hear ye! I totally agree with you. I've always thought about a no-TO FFL, built on somewhat simplistic notions derived from "Hidden Game of Football" and FO. I.e. Divide all yardage by 12 points; add up all turnovers and 4th down misses + missed field goals. It seems somewhat insane to most FFL players; maybe my view is heavily skewed because I'm totally pissed about Ray Rice vs. Willis McGahee usage patterns.

88
by tuluse :: Wed, 10/07/2009 - 2:20pm

What you guys propose is very boring.

Fantasy football is fun because TDs are so random.

What you guys want to do is just rank the best players.

92
by Samoan_Rob :: Fri, 10/09/2009 - 9:50am

By that logic the NFL is boring because all it wants to do is pick the best team

32
by jimm (not verified) :: Tue, 10/06/2009 - 2:27pm

Will - There sure didn't seem to be any holes for Peterson in the 2nd half but Childress did comment that he was banged up a little after the fumble. Didn't look the same to me in the 2nd half and clearly looked groggy after the fumble play.

34
by George (not verified) :: Tue, 10/06/2009 - 2:39pm

"At what point does he become the real Steve Smith and the Panthers one becomes the other guy?" When Eli Manning gets traded for Jake Delhomme and the Giants Smith still outperforms Carolina's Smith, then I'll buy this.

37
by countertorque :: Tue, 10/06/2009 - 2:56pm

I know the defensive adjustments are only at 40% right now. But, citing the quality of the defense in your arguments for why players scored so high in your defense adjusted metrics seems strange.

42
by Eddo :: Tue, 10/06/2009 - 3:09pm

Why? If the defensive adjustment (as opposed to fumble luck or not converting first downs) is what creates the discrepancy between conventional statistics and DYAR, then it's worth noting.

45
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 3:18pm

Nono, he's talking about the Matt Forte comment: Forte was the top running back by DYAR, and the claim is because it was against the Lions. But defensive adjustments should take that away, and normalize it to around his previous outings. So the comment seems odd.

I don't think it's because defensive adjustments are only at 40%, though. It's because we're still in small-sample size land; the Lions rush defense is actually above average, but that's almost certainly because their pass defense has been god-awful.

50
by Eddo :: Tue, 10/06/2009 - 3:44pm

Gotcha. I misinterpreted the original comment.

51
by Bill Barnwell :: Tue, 10/06/2009 - 3:52pm

Sorry for the confusion. What I was trying to say was:

a) Matt Forte had a great day
b) But it came against the Lions
c) The day was great, still, but when opponent adjustments get to 100%, it's not going to look as good as it does now
d) At that point, his first three weeks (particularly Week 2) will likely look a lot better, and Week 4 will look way worse.

64
by Pat (filler) (not verified) :: Tue, 10/06/2009 - 4:56pm

Bill:

That argument would be a lot more sound if the Lions actually had a bad run defense DVOA. They don't. Through week 3, they were actually above average.

Now, I believe that they probably will have a below-average run defense eventually, because it was almost certainly selection effect from weeks 1-3, but hey, who knows.

52
by goforit (not verified) :: Tue, 10/06/2009 - 3:54pm

is it just me or are teams going for it on 4th down a lot more than normal?

55
by Bobman :: Tue, 10/06/2009 - 4:01pm

I hope it's a trend and not just you.

67
by Kellerman :: Tue, 10/06/2009 - 5:10pm

The "Blood Pressure" Bengals are certainly going for it on 4th down repeatedly, and have made nearly every one of them whether its 4th and 3 and a pass to Coles or 4th and goal and passed to Chad or at least two 4th and 10s one of which converted by passing and one by Palmer scrambling. They've also succeeded on a fake punt.

57
by Oscar (not verified) :: Tue, 10/06/2009 - 4:20pm

Watching Aaron Rodgers last night made me realize how great Roethlisberger is.

59
by Grouchy Bills Fan (not verified) :: Tue, 10/06/2009 - 4:42pm

*&*(##! Trent Edwards can't even succeed at being the worst.

69
by Kellerman :: Tue, 10/06/2009 - 5:18pm

This may be more appropriate for the DVOA rankings thread but I've noticed that DVOA confirms my impression that the Bengals have completely shut down the opponent's #1 WR so far to the tune of -99% and that's before holding Braylon Edwards to no catches in week 4. My impression is that Leon Hall is playing in Asomughalike fashion thus far. Jonathon Joseph appears to be being picked on and this is supported by the Bengals' #30 ranking against #2 WR even before Mohamed Massaquoi had his breakout game this week.

Does charting support the notion of Hall's excellence or is he splitting time on the enemy #1s? Or is it just a schematic thing to take away the #1 and say "if the others can beat us then so be it?"

87
by Oldcat (not verified) :: Wed, 10/07/2009 - 2:14pm

Hall has been playing well, but a good fraction of Massaqoi's catches came against him last week.