Writers of Pro Football Prospectus 2008

Most Recent FO Features

McGloinMat13.jpg

» Week 4 Quick Reads

Our look at which quarterbacks were "streakiest" in Week 4 includes more discussion of Raiders quarterbacks than we ever anticipated.

24 Dec 2012

Week 16 Quick Reads

by Vince Verhei and Aaron Schatz

Detroit Lions wide receiver Calvin Johnson made history this weekend against Atlanta. Johnson gained 225 receiving yards against the Falcons, bringing his season total to 1,892, breaking the record of 1,848 set by San Francisco's Jerry Rice in 1995. That record lasted for 17 years. Eric Dickerson's record of 2,105 yards rushing, set in 1984, has lasted even longer, but that mark too is in jeopardy. Despite a modest output on Sunday against Houston, Minnesota's Adrian Peterson has 1,898 yards on the ground this season, and he can surpass Dickerson's mark with 208 yards next week against Green Bay -- the same Green Bay team that let Peterson run for 210 yards in Week 12.

There's a problem with all these numbers, though. Total yardage only tells us part of what each player did in a given season. In a nutshell, it tells us a lot about the good plays a runner or receiver made, but it doesn't say much about the bad plays -- the incomplete passes, the stuffs in the backfield, the fumbles, and others. For that we can turn to DYAR (Defense-Adjusted Yards Above Replacement), Football Outsiders' metric that examines every play of the NFL season and measures its value in terms of producing yards, first downs, turnovers, and touchdowns, then adjusts for factors such as down, distance, field position, score, opponent, and other factors. According to DYAR, Johnson's season has been truly elite, but Peterson's campaign has been less valuable to the Vikings than fantasy numbers or highlight reels would have you believe.

We'll start with Johnson, since his status is more straightforward. The following table shows the top wide receiver seasons in DYAR since 1991, including Johnson this season (projected over 16 games):

Top Wide Receiver Seasons by DYAR, 1991-2012

Year Name Team DYAR Passes Catches Yards TD C% Yds/Pass
1995 Michael Irvin DAL 599 165 111 1,603 10 67% 9.72
2011 Calvin Johnson DET 570 158 96 1,680 16 61% 10.63
2007 Randy Moss NE 564 160 98 1,482 23 61% 9.26
2001 Marvin Harrison IND 548 164 109 1,524 15 66% 9.29
2012 Calvin Johnson DET 533* 203* 125* 2,018* 5* 62% 9.96
2011 Jordy Nelson GB 517 96 68 1,263 15 71% 13.16
2006 Marvin Harrison IND 513 148 95 1,366 12 64% 9.23
1995 Jerry Rice SF 512 175 122 1,848 15 70% 10.56
1994 Jerry Rice SF 512 150 112 1,499 13 75% 9.99
2008 Andre Johnson HOU 510 171 115 1,575 8 67% 9.21
2003 Randy Moss MIN 508 172 112 1,632 17 65% 9.49
* Current statistics projected over 16 games.

Johnson's catch rate and yards per pass are good, but nothing special among this group, and his touchdown total is relatively poor. Johnson's most eye-popping statistic (besides his yardage, obviously) is the sum of his targets. Johnson has 190 targets so far this season with one game to go. He's a safe bet to become the fourth player since 1991 to amass 200 targets in a season, joining Herman Moore (who had 206 for the 1995 Lions), Marvin Harrison (202 for the 2002 Colts) and Rob Moore (208 for the 1997 Cardinals). Johnson is delivering like nobody else ever has in history, but that's largely because he's been given opportunities to deliver like very few men before him.

By the way, if Johnson goes wild against the Bears this Sunday, he has a very slim chance to finish atop this list. He'll need 100 DYAR against Chicago to top Michael Irvin's 1995 season for Dallas. The best game for a receiver this year was Andre Johnson's 85-DYAR outing against Jacksonville in Week 11. The Houston Texans wideout caught 14 passes in 19 targets for 273 yards and a touchdown against the Jaguars. That's the ballpark of what Calvin Johnson will need this weekend, though he'll also get a sturdy boost in DYAR for playing the mighty Chicago defense.

(Ed. Note: You may be wondering how Michael Irvin had more DYAR than Rice in 1995 despite having fewer yards, fewer touchdowns, and a lower catch rate. The answer has to do with three things: better third-down performance, two fewer fumbles, and 150 additional yards from pass interference calls.)

Now for Peterson. The Vikings running back does not project to finish among the top runners in our DYAR database, but we're listing him here with the leaders to demonstrate what his strengths and weaknesses have been this season, and how extreme those strengths and weaknesses have been.

Top RB Seasons By Rushing DYAR, 1991-2012
Year Name Team DYAR Runs Yards Avg. TD FUM Success
Rate
Stuff
Rate
1998 Terrell Davis DEN 602 392 2,008 5.12 21 1 52% 18%
1999 Stephen Davis WAS 526 290 1,407 4.85 17 3 60% 12%
1997 Terrell Davis DEN 526 369 1,743 4.72 15 4 56% 13%
1995 Emmitt Smith DAL 505 375 1,770 4.72 25 7 53% 17%
2000 Marshall Faulk STL 501 253 1,359 5.37 18 0 61% 11%
2002 Priest Holmes KC 497 313 1,615 5.16 21 1 55% 18%
2005 Larry Johnson KC 488 335 1,741 5.20 20 5 55% 14%
2003 Priest Holmes KC 485 320 1,420 4.44 27 1 58% 18%
1994 Emmitt Smith DAL 461 368 1,484 4.03 21 1 52% 17%
2006 LaDainian Tomlinson SD 460 347 1,813 5.22 28 2 49% 14%
2012 Adrian Peterson MIN 388* 335* 2,025* 6.04 12* 3* 49% 24%
* Current statistics projected over 16 games.

Peterson's carry and fumble numbers sit in the middle of this pack, and though his touchdowns are on the low side, he more than makes up for it with an average gain that blows these players – the best of the best – out of the water. However, his success rate pales in comparison to the rest of his peers, and he gets stuffed in the backfield at a staggering rate, more than twice as often as some of these other backs. Those negative runs leave Christian Ponder and the Vikings offense in long-yardage situations far too often, and that's why Peterson's DYAR will fall outside the top 10 in the DYAR database. With 388 DYAR, Peterson would rank 17th.

(By the way, this high-average/high-stuff-rate combo is not unprecedented. In 1997, Detroit's Barry Sanders gained 2,053 yards and averaged 6.13 yards per carry, but he was stuffed on 22 percent of his runs.)

Is Peterson solely responsible for his stuff rate? Of course not – his offensive linemen deserve their share of the blame as well. On the other hand, they also deserve some credit for clearing the path for Peterson to rip off some of his longer runs. Our in-depth play-by-play analysis gives us a lot more information than simple yardage totals, but we still can't totally separate the performance of a runner from his offensive line. For what it's worth, Peterson's teammate Toby Gerhart has a stuff rate of 22 percent, which is very high, but still lower than Peterson's, and Minnesota's offensive line fares well in our run-blocking metrics.

If this sounds like an attempt to disparage Johnson and (especially) Peterson, that's not the intent. Over the course of the season, our numbers say they have still been the most valuable receiver and runner in the league this year. There's a lot more to winning games than racking up real estate, though, and as great as this duo has been, we have seen a handful of more valuable players in the past couple of decades.

Quarterbacks
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
1.
Drew Brees NO
37/53
446
3
0
232
232
0
Fifty-three passes, but only three deep balls, two of which he completed for a total of 47 yards. On third downs, he went 12-of-16 for 183 with nine first downs, plus a touchdown, and a 3-yard DPI. This currently stands as the third-highest QB DYAR of the season, behind Aaron Rodgers (Week 6 vs. HOU) and Tom Brady (Week 8 vs. STL).
2.
Joe Flacco BAL
25/36
318
2
0
196
205
-10
In his last five games, Flacco completed 57 percent of his passes with 6.4 yards per pass. That kind of production from the quarterback position isn't going to earn many wins against playoff teams. The Ravens' game against the Giants on Sunday had serious postseason ramifications, and a Baltimore loss would have been their fourth in a row, and set up a winner-take-all showdown for the AFC North title in Week 17. Flacco showed up big against New York, though. He was at his best on third downs, going 11-of-14 for 137 yards and 10 first downs, including a goal-to-go touchdown, plus a 17-yard DPI. He even had a pair of conversions on separate third-and-19 plays.
3.
Matt Stafford DET
37/56
443
0
1
186
193
-8
On passes to the middle of the field, Stafford went 12-of-13 for 213 yards and nine first downs, and he would have had a tenth first down if Calvin Johnson hadn't fumbled the ball away.
4.
Matt Ryan ATL
25/32
279
4
0
179
167
12
There's a pattern forming with Ryan where he is at his best early and late in games. He hit each of his first 12 passes for 156 yards with two touchdowns and three other first downs. He may have been even better in the fourth quarter, where he had a stretch of eight completions in a row for 77 yards with one touchdown and six other first downs.
5.
Aaron Rodgers GB
27/38
342
3
0
165
167
-2
First half: 15-of-25 for 152 yards with one sack, one touchdown, and six other first downs. Second half: 12-of-13 for 190 yards with two touchdowns and nine other first downs.
6.
Peyton Manning DEN
30/43
339
3
1
164
164
0
Third downs: 8-of-12 for 122 yards with two touchdowns, six other first downs, and one interception.
7.
Tony Romo DAL
26/42
416
4
0
161
161
0
On deep passes, Romo went 6-of-10 for 176 yards and three touchdowns.
8.
Russell Wilson SEA
15/21
171
4
1
128
112
16
Red zone passing: 5-of-6 for 36 yards with four touchdowns and a first down on third-and-6.
9.
Jay Cutler CHI
12/26
146
1
0
84
78
6
Well, here's a switch. This week, it's the Bears' own quarterback who gets a boost for playing a good defense, and not just their opponent. He only had one touchdown and five other first downs against Arizona, and those six plays picked up 110 yards. Otherwise, he went 6-of-20 for 36 yards with a sack and eight failed third-down plays with less than ten yards to go for a first. Also, Cutler continues to be the worst first-quarter quarterback in football, going 0-for-6 with a sack. His first-quarter statline this year: 51-of-101 for 395 yards with three touchdowns, six interceptions, 11 sacks, and a jaw-dropping -420 DYAR. No other quarterback is below -196 DYAR in the first quarter.
10.
Christian Ponder MIN
16/30
174
1
0
55
44
12
Ponder also gets a healthy boost for playing a good defense. He started off 3-of-3 for 62 yards with a touchdown and two other first downs, but had only five more first downs the rest of the day. On Houston's half of the field, he went 7-of-15 for 75 yards and only four first downs, including the touchdown.
11.
Colin Kaepernick SF
19/36
244
1
1
41
41
0
Another quarterback who gets a boost from playing a strong defense, Kaepernick also had some heavy duty stats padding. His first pass inside Seattle's 40 was an 11-yard gain on third-and-11, but he didn't pick up another first down in that territory until San Francisco trailed by 29 points in the third quarter. His final numbers inside the 40: 9-of-19 for 59 yards with one interception, one sack, one touchdown, and four other first down. (And it's not his fault, but his receiver also fumbled the ball away on one of those completions.) Three of those first downs, including the touchdown, came when the 49ers trailed by 36 points in the fourth quarter.
12.
Philip Rivers SD
11/22
165
2
0
36
37
-1
Rivers didn't have much luck once he crossed his own 40. In the Red, Front, and Mid zones, he went 6-of-13 for 109 yards with three sacks and only four first downs, although two of those first downs were touchdowns of 34 and 37 yards.
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
13.
Sam Bradford STL
13/27
202
2
1
32
32
0
First four plays of the second half: 80-yard touchdown, 8-yard gain on third-and-3, 11-yard gain on first-and-10, 5-yard touchdown on second-and-goal. When he got the ball back, the Rams were up 15 in the third quarter. Bradford proceeded to throw six incompletions in a row. His last pass of the day was finally a completion ... for a 4-yard loss on third-and-12.
14.
Robert Griffin WAS
16/24
198
2
1
26
26
1
Griffin threw touchdowns of 22 and 11 yards against Philadelphia, but had just one other first down inside the Eagles' 40. His final numbers in that range: 4-of-11 for 56 yards.
15.
Colt McCoy CLE
9/17
79
1
0
20
10
10
Football Outsiders alumnus Bill Barnwell on Twitter: "Colt McCoy is one-hopping passes literally five yards short of his receivers. Repeatedly." And he still finished higher than Brandon Weeden. Twenty-one of McCoy's 22 plays came in the fourth quarter, in three distinct stretches. He started out 3-of-7 for 14 yards with a sack and no first downs. He then completed five passes in a row for 61 yards, with every completion gaining a first down or touchdown, and also had a 19-yard DPI in that stretch. And then he finished 1-of-5 for 4 yards with no first downs and three sacks.
16.
Brandon Weeden CLE
12/19
104
0
0
18
15
2
On Denver's side of the field, Weeden went 3-of-6 for 25 yards with two first downs and a sack.
17.
Chad Henne JAC
29/49
348
1
3
8
-4
12
In a game the Jaguars lost by seven points, Henne had many chances late to be a hero. He had six red-zone throws in the fourth quarter and produced two completions for 14 yards and no first downs, one incompletion, one sack, and two interceptions (although one of those interceptions came on fourth down and the other, with 5 seconds to go, is treated like a Hail Mary). He did throw a red-zone touchdown in the first quarter, but he also had an incompletion on third-and-6 at the New England 17 in the second. The Jags kicked a field goal there. If they had scored a touchdown, they would have been in position to tie the game with a field goal late in the fourth.
18.
Ryan Lindley ARI
17/30
141
0
1
6
6
0
Third downs: 5-of-8 for 27 yards, one sack, and only two first downs.
19.
Nick Foles PHI
32/48
345
1
1
3
4
-1
Red-zone passing: 3-of-9 for 27 yards, two first downs, no touchdowns, and two sacks. Keep in mind the Eagles lost this game by seven points.
20.
Eli Manning NYG
14/28
150
1
0
2
-1
3
Third downs: 4-of-8 for 31 yards with two first downs and a sack. He didn't convert a third-down play until the Giants were down by 20 points in the third quarter.
21.
Ryan Tannehill MIA
13/25
130
2
0
-6
-5
-1
So much for the element of surprise. On first downs, Tannehill went 3-of-9 for 31 yards with one sack and only one first down, although that first down was a 17-yard touchdown.
22.
Andrew Luck IND
17/35
205
1
0
-10
-4
-5
In one stretch of the second half, Luck went 0-for-10 with a sack. Then he finished 6-of-7 for 49 yards with five first downs, including the game-winning touchdown.
Rk
Player
Team
CP/AT
Yds
TD
INT
Total
DYAR
Pass
DYAR
Rush
DYAR
23.
Tom Brady NE
24/41
267
2
2
-21
-21
0
A very un-Brady like performance on third downs: 3-of-10 for 29 yards with one sack and three first downs, though one of those was a touchdown on third-and-goal from the 2.
24.
Andy Dalton CIN
25/41
278
0
2
-26
-26
0
Inside the Pittsburgh 40, Dalton went 0-for-5 with a sack and an intentional grounding.
25.
Cam Newton CAR
18/29
170
1
1
-26
-44
18
On Oakland's side of the field, Newton went 5-of-8 for 67 yards, but only two first downs, although one of those was a 23-yard touchdown. Also had eight carries for 61 yards, including gains of 19 and 29 yards, and a touchdown on third-and-goal from the 3.
26.
Matt Schaub HOU
18/32
178
0
0
-48
-48
0
Third downs: 4-of-8 for 22 yards with two sacks and only one first down, plus a 12-yard DPI.
27.
Brian Hoyer ARI
11/19
105
0
1
-51
-51
0
On Chicago's half of the field, Hoyer went 3-of-7 for 16 yards with one first down, two sacks, and one fumble.
28.
Ryan Fitzpatrick BUF
20/35
240
1
1
-54
-61
8
Third downs, first three quarters: 2-of-4 for 8 yards with no first downs and a sack-fumble. By then Buffalo was down by 21 points. In the fourth quarter, he went 5-of-5 on third downs for 76 yards and four first downs.
29.
Matt Leinart OAK
16/32
115
0
1
-67
-67
0
Third and fourth downs: 8-of-15 for 53 yards with one interception, one sack, and four first downs.
30.
Ben Roethlisberger PIT
14/28
223
1
2
-91
-87
-4
Third downs: 1-of-8 for 13 yards with a pick-six, three sacks, and a fumble. Hey, at least that one completion went for a first down.
31.
Brady Quinn KC
10/22
162
0
2
-93
-83
-10
On Indianapolis' half of the field, Quinn went 3-of-9 for 39 yards with one first down, one sack, and one interception.
32.
Greg McElroy NYJ
14/24
185
0
1
-109
-114
5
Not shown in the numbers of this table: McElroy's 11 sacks. Twice he was sacked on back-to-back-to-back dropbacks. He was sacked twice on first down, four times on second, four times on third, and once on fourth. He was sacked once in the first quarter, four times in the second, twice in the third, and four times in the fourth. Two sacks came in the Deep zone (inside the Jets' own 20), seven in the Back zone (between the Jets' 20 and 40), and two in the Mid zone (between the 40s). His average sack came with 11.3 yards to go for a first down, and resulted in a loss of 6.2 yards.
33.
Josh Freeman TB
30/54
372
1
4
-122
-127
5
Inside the Rams' 40, Freeman went 7-of-17 for 62 yards with two sacks and only three first downs.
34.
Jake Locker TEN
14/30
140
1
2
-152
-162
9
Locker's first first down was an 8-yard gain on third-and-7 in the second quarter. By that point he had already given up two sacks, two interceptions, and an intentional grounding, and the Titans were behind by 17 points. He was sacked seven times on the day. He only had seven first downs all day, and three of them (including his touchdown) came in the game's final two minutes with Tennessee down by 55. On second downs, he went 1-of-6 for 39 yards (that completion also came down 55 points in the fourth) with an intentional grounding and four sacks.


Five most valuable running backs
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
Marshawn Lynch SEA
111
1
19
1
62
34
28
The Seattle Seahawks have scored 150 points in their past three games, thanks in large part to Lynch's production. Lynch has been the top running back in our rankings for three weeks in a row now, which is quite possibly a first in the history of Quick Reads. He's only carrying the ball about 15 times a game in that stretch, but he's averaging 7.5 yards per carry. He has been stuffed for no gain or a loss four times, but he's run for 10 or more yards 12 times, with five touchdowns and 10 other first downs. He only has three receptions in those three games, but each of those catches has gained at least 9 yards and a first down, including a touchdown on second-and-goal from the 9.
2.
LeSean McCoy PHI
45
0
77
0
57
14
43
Between runs and receptions, McCoy had nine first downs, including a must-have fourth-and-2 near the end of the game.
3.
Darren Sproles NO
48
0
104
0
47
12
35
It was a big week for running backs in the receiving game. Sproles had five different receptions for at least 10 yards
4.
Montell Owens JAC
42
0
77
0
46
7
39
Before four weeks ago, Montell Owens had 83 career rushing yards, half of those on one fake punt back in 2008. He has 209 rushing yards since then. He also caught all four of his intended passes this week for 77 yards, including taking a pass behind the line of scrimmage 53 yards in the first half.
5.
Ryan Grant GB
80
2
34
0
42
31
11
Hey, remember this guy? Grant was the same Grant he always was, with few breakaway runs (his only double-digit run went for 18 yards) but a lot of consistent gains (16 of his 20 runs gained at least wo yards). He converted twice on fourth-and-short, and again at the goal line.


Least valuable running back
Rk
Player
Team
Rush
Yds
Rush
TD
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
1.
BenJarvus Green-Ellis CIN
14
0
0
0
-48
-42
-5
Oof. Fifteen runs for only 14 yards. The Steelers defense stopped Green-Ellis for a yard or less on nine out of those 15 runs, and he didn't have a single first-down conversion. Also poor this week: Arian Foster at -30 DYAR (10 runs for 15 yards) and Darren McFadden at -34 DYAR (17 runs for 33 yards).


Five most valuable wide receivers and tight ends
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Dez Bryant DAL
9
12
224
24.9
2
83
Did you look up at your television set and see "Dez Bryant: 58-YARD TD RECEPTION" on the scroll and think, "Oh, yeah, they said that a few minutes ago?" Well, that was actually a different play. Bryant had twin 58-yard touchdown receptions. He would have challenged Andre Johnson for the best wide receiver game all season if not for the fact that he only caught one of four targets on third down.
2.
Calvin Johnson DET
11
16
225
20.5
0
76
Megatron had nine first downs against one of the league's better pass defenses, but his DYAR was hurt by the ball he fumbled away in the second quarter.
3.
Roddy White ATL
8
10
153
19.1
2
74
You don't often see Roddy White racking up the YAC, because he's usually a deep threat, but in the second quarter he caught a pass near the line of scrimmage on a second-and-5 and ended up with a 39-yard touchdown.
4.
Eric Decker DEN
6
6
65
10.8
2
63
Only one of the six passes to Decker didn't convert for a first down or a touchdown; that one went nine yards on first-and-10.
5.
Larry Fitzgerald ARI
8
12
111
13.9
0
46
Oh my god, he's still alive! Surprisingly, Fitzgerald was worth more this week when Ryan Lindley was in the game than when Lindley was replaced by Brian Hoyer. Lindley completed four of five passes to Fitzgerald for 61 yards, plus a 10-yard DPI. Hoyer connected with Fitzgerald on four of seven passes for 50 yards, and threw an interception on a pass meant for Fitzgerald. (We don't penalize Fitzgerald himself for that.)


Least valuable wide receiver or tight end
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
1.
Danny Amendola STL
2
7
11
5.5
0
-51
Amendola caught only two of seven passes, and fumbled one of those away, against a Tampa Bay team that ranks just 28th against the pass. Also poor this week: Steve Johnson caught just 4 of 10 passes for 33 yards (-43 DYAR), and DeVier Posey caught just one pass (out of six) for six yards (-35 DYAR).

Posted by: Vincent Verhei on 24 Dec 2012

37 comments, Last at 01 Jan 2013, 1:29pm by Noah of Arkadia

Comments

1
by tally :: Mon, 12/24/2012 - 7:57pm

Maybe I missed this over the years, but is there a reason why FO doesn't choose to report median YPC instead of mean YPC? It would help encapsulate both average yardage and success rates while reducing the influence of long carries and probably better correlate to DVOA and DYAR.

6
by An Onymous (not verified) :: Tue, 12/25/2012 - 12:09am

Because the median ypc of every RB in history 2. Seriously, pretty much every RB to play the game has a median carry of 2 yards. It'd be more informative to know the IQR, but really, a combination of ypc and success rate is probably the most efficient way to communicate all of the relevant information in the least amount of time.

8
by RickD :: Tue, 12/25/2012 - 12:45am

You think the median ypc is 2 yards? For every RB in history?

So, you're saying half of all running plays, for all RBs out there, gain 2 yards or less.

I would disagree.

(googles)

This post decisively disproves your contention by looking at the percentage of carries of at least 4 yards for several RBs.

http://www.advancednflstats.com/2007/08/median-rushing-yards.html

Stuff rate relates to this, of course, as does success rate. But success rate is less useful, since a short yardage RB could have a high success rate while still only getting 1-2 yards per attempt.

I suspect most successful RBs have median gains of at least 3 yards. It's hard to average 4+ yards/carry if more than half of your runs are less than 2 yards.

2
by Ben :: Mon, 12/24/2012 - 8:53pm

How did the Chiefs RB's rank before defensive adjustments? Was it the Colt's defensive modifier that dragged them out of the top 5?

3
by Zach (not verified) :: Mon, 12/24/2012 - 9:18pm

To be fair, I think Kaepernick actually does deserves some of the blame for Manningham's fumble (and injury), as he took FOREVER to get the ball out to him on that screen play.

4
by commissionerleaf :: Mon, 12/24/2012 - 9:41pm

Absolutely correct. Kaepernick was about to throw it and pumped, I think because Aldon Smith was jumping into the air to get a hand on the screen pass. Kaepernick paused and pulled down the ball, and leaned back to throw around Smith. As a result, two defenders were waiting to plaster Manningham. Kaepernick either needed to lob over Smith, get the pass off before Smith was prepared to jump, or bail out and throw somewhere else.

10
by Zach (not verified) :: Tue, 12/25/2012 - 2:30am

It was actually Chris Clemons (because Smith plays for the 49ers), but yes, that's what happened.

15
by commissionerleaf :: Tue, 12/25/2012 - 1:01pm

Of course it was. I remember thinking Clemons while watching it, but then blanked writing the post and just thought about who mans the right side without thinking about team.

Boo me. Thank you.

5
by greybeard :: Mon, 12/24/2012 - 10:04pm

Every week you have to write why a QB who played a terrible or below average game is near the top because they played a good defense. It is especially interesting that bears defense apparently plays a top QB evert week regardless of who the QB is. I am exaggerating of course, but seems to me you can improve how you calculate the DYAR for QBs by two changes:
1) adjustment for opponent should be dialed down
2) garbage time stat padding should be weighted less.

Also, nice effort to put Johnson's number in context by using DYAR. However, are there anybody besides FO staff that think that DYAR is working for RBs or WRs? I think traditional numbers are much better representative of success for these positions.
I like what pro-football-reference guys do. They adjust for era and stay there.

7
by An Onymous (not verified) :: Tue, 12/25/2012 - 12:11am

I think DYAR is pretty good for RBs, but I agree that it's woeful for measuring WRs. You see a ton of #2 WRs getting huge DVOA totals because they're facing junk coverages.

20
by Kal :: Tue, 12/25/2012 - 4:31pm

It's pretty horrible for RBs because its meant to be predictive, not descriptive. So jamaal Charles gets 226
Yards on 20 carries and doesn't break the top 5 dyar because after 20 yards long runs aren't predictive. Okay, but they're hugely valuable. Same goes for AP most weeks. Or forte last year when he was the bears offense one game and didn't get into the top 5 despite having over 200 yards.

In general FO overvalues long sustained drives and undervalued explosive drives. This especially shows up when evaluating running backs.

21
by Noah of Arkadia :: Tue, 12/25/2012 - 7:55pm

There's a very good reason for that. Offenses score more points through long drives than through big plays. Hence long drives are more valuable. You mentioned 20 yards. Which offense has more TDs over 20 yards than TDs under 20 yards? I bet none. Ever.

------
FO posters are a peacock. You got to let us fly!

23
by Kal :: Wed, 12/26/2012 - 12:17pm

That's a false argument. You don't declare something worthless because something else is more common. It doesn't matter if long drives are the norm. What matters is how valuable that play was. For the Vikings, for instance, AP is often the only offense they have - and that means they are not going to score very often on long drives, making those bursts even more important.

It's fine (though probably flawed) to say that it is a predictive stat and that past 20 yards it loses any predictive value. Okay. I actually suspect that this is incorrect for RBs now, but lets go with it. That doesn't mean that you can look at dyar and say that it is in turn particularly descriptive. A player with 2 long td runs and 10 stuffs is a hell of a lot more valuable and important to their teams success than a player with 12 plays that succeeded but scored nothing or gained only 60 yards, but that's what dyar would rate it as.

24
by JoeyHarringtonsPiano :: Wed, 12/26/2012 - 12:50pm

What you're forgetting to mention is that the first player you mentioned likely killed his team's drives on those 10 stuffs, giving the opposing team good field position. The second player, even if he didn't break off a TD run, at worst allowed his team to flip field position, making it less likely for the opposing team to score, and at best put his team in the red zone so another player on his offense can score a TD.

27
by Kal :: Wed, 12/26/2012 - 5:38pm

That's possible. Another alternative is that the first player simply got stuffed and had longer to run when breaking that giant run, making two of their stuffs completely moot. Or the defense got a 3 and out on the punt, and thanks to their great punting and defense they've been doing that all year. I mean, you don't say that a QB 'killed his team' when throwing one incomplete pass (or 10, for that matter), right?

Ultimately you have to decide what was actually more valuable for that game, not theoretically more valuable. DVOA is predicated on success rates being the most important thing, with explosiveness not mattering nearly as much (and being devalued quite heavily). That's fine - but that doesn't make it perfect. S+P rates explosiveness significantly higher, for instance. FEI doesn't care about either so much as it cares about drive (not play) effectiveness - so in FEI, you'd see whether or not in a game those 10 stuffs actually mattered or not.

Here's the eye test: do you think that Jamaal Charles was less valuable to his team on Sunday than, say, Montel Owens? Do you think AP was and is less valuable to his team in week 15 than DeAngelo Williams, he of 22 carries on 93 yards + 60 yards receiving on two catches? If you do, that's fine - but I think a very reasonable argument can be made that this is missing out on significant value because of DVOA's inherent biases, and often does not remotely pass the eye test when evaluating running backs.

28
by LionInAZ :: Thu, 12/27/2012 - 3:04pm

If said QB threw 10 incompletions on 3rd down, yes, you would say that he 'killed his team'.

31
by Kal :: Fri, 12/28/2012 - 9:49pm

That's a false equivalence; no one is comparing 10 3rd down runs that were stuffed. And point of fact, from what I gathered DVOA rates being stuffed on 1st as a negative value just as much as 3rd, sometimes more so because it's rarer. Again, another flaw.

29
by Noah of Arkadia :: Thu, 12/27/2012 - 3:46pm

Exactly. Generally speaking, in any team sport the team that plays as a team, with players setting each other up for success, will perform better than the team that relies on the talent of its individuals to make big plays. That's what makes the lowly ten 4-yard carries more valuable than the 40-yard carry and nine stuffs.

Kal, you didn't like my argument, but working as a team -just doing your job every play- leads to long successful drives, which is what you need to win in this league. 80-yard runs are very valuable indeed, but their value is diminished if they come at the cost of crippling even two or three other drives.

------
FO posters are a peacock. You got to let us fly!

30
by Kal :: Fri, 12/28/2012 - 9:48pm

That's the thing - it's not a zero sum game. You can have 80 yard runs and have a few stuffs and the two do not negate each other. They don't cancel each other out. Why would they? Does throwing an 80 yard bomb hhave the same equivalence value as 3 incomplete passes?

Long, successful drives are not the only way teams score. DVOA rewards teams for doing it that way because that is more predictive of future success, but that doesn't mean it's more valuable for that specific game. I agree that everyone doing their job well can lead to long successful drives. So what? Everyone doing their job can also lead to being crushed horribly. Sometimes it's just fine to be lucky in a game and you can win that way happily - and more importantly that's a fine descriptor of how that specific game was played.

The problem is that DYAR is more about the prediction. And that's fair, but let's not say that players like Charles and AP played worse than other players out there; they simply are not likely to be successful like that on a consistent basis. That isn't necessarily bad. It's just not predictive. Put it another way: would you rather have a back like Lynch who is fairly consistent at getting a few yards on Minnesota's offense, or would you rather have AP?

Consistency is not the be-all, end-all of football. Explosiveness is important as well. Blind adherence to DVOA biases does no one favors - especially those on this site.

32
by Purds :: Sat, 12/29/2012 - 12:52pm

I know I am very late to the party, but I completely agree with you on these points, Kai. DVOA is fine for predicting future success but is not really an accurate read on how much a player helped his team in the past game(s). As the equivalent of AP, I offer Andrew Luck. He is not performing well via DVOA and DYAR stats, and as a Colt fan I do worry about his long-term potential for success, as those stats indicate I should. However, they don't do a very good job of realizing that Indy, as currently staffed, will have little to no success trying to keep sustained drives alive (the favorite of the DYAR stat). Indy must rely on the longer plays, the downfield throws, and thus Luck has some pretty bad DVOA/DYAR days even though they have won some games because of, as you describe, explosive plays. Yes, I completely agree with DYAR and DVOA that Luck will not have future success playing like this, but I would offer that Luck has helped his team this year more than, say for example, Carson Palmer, who is a tad above him in rank but with almost twice the DYAR. Really? Carson Palmer has been twice as successful as Luck this year?

33
by LionInAZ :: Sat, 12/29/2012 - 5:27pm

I'm just shaking my head.

It's all about what is going to score the most points. An 80-yd run gets you at most one score. 8 first downs gives you a chance to get 2 scores.

Repeatedly getting stuffed leaves you in 2nd-and-long and 3rd-and-long situations a lot more, and also puts your defense on the field more. Do you really think the occasional long run negates all that?

There's a reason the Vikings score only 2 offense TDs per game.

34
by Kal :: Sat, 12/29/2012 - 8:43pm

If you really believe that Adrian Peterson is the reason that the Vikings only score two TDs a game and they'd be better off with someone else, I'm not sure what you're on. The reason the Vikings score so little is because their offense is essentially AP and nothing else. They're totally anemic otherwise. Their passing game is horrible. And you blame that on AP? Come on.

An 80 yard run vs. 8 consecutive 1st downs gives you the same chance to score if they're on the same drive. The real problem is that an 80 yard run isn't being compared to 8 first downs. It's being compared to about 4 successful plays, one of which is a first down. That means, roughly, about 20 yards and successful runs on each. If it were the case that 8 first downs were the same value as an 80-yard run, that'd be one thing; it isn't.

As an example, the game last year with the Pats/Jets where the Jets scored more on DVOA than the Pats despite being beat by 14 points. The reason was because the Jets drives were either 3 and outs or 12-15 plays where every play was a success. They scored 3 times. This was marked as an AMAZING DVOA game. Why? Because DVOA does averages and not drives - so if the Jets had 8 drives, they would have had at most 15 unsuccessful plays and 40 successful ones, making them look great. Turns out that's not really realistic either.

Repeatedly getting stuffed leaves you in 2nd-and-long and 3rd-and-long situations a lot more, and also puts your defense on the field more. Do you really think the occasional long run negates all that? Depends a lot on the other situations and whatnot, but often yes. I agree that it isn't predictive, but I would much rather take a 80 yard TD run and a few stuffs over 4-5 successful plays from my RB that net about 25 yards and keep me 'in the count'.

35
by Andrew Potter :: Sat, 12/29/2012 - 10:22pm

If you really believe that Adrian Peterson is the reason that the Vikings only score two TDs a game and they'd be better off with someone else, I'm not sure what you're on. The reason the Vikings score so little is because their offense is essentially AP and nothing else. They're totally anemic otherwise. Their passing game is horrible. And you blame that on AP? Come on.

FO's metrics don't believe that the Vikings would be better off with another back either - Adrian Peterson leads the league in DYAR and is second in DVOA. He doesn't seem to be a particularly good example of either metric being flawed.

36
by LionInAZ :: Sun, 12/30/2012 - 5:19pm

I didn'T say anything about getting 8 consecutive first downs. But getting 8 first downs makes it more likely your team will get multiple scores than one 80 yd run. Why is this so difficult to understand?

Peterson is certainly not the main reason the Vikings offense is bad. But he's not keeping them out of a lot of obvious passing situations either. That's the point!

The same arguments have been made to downgrade Barry Sanders' career.

37
by Noah of Arkadia :: Tue, 01/01/2013 - 1:29pm

Exactly. This is not about AP not being valuable. It's about how those stuffs counterbalance the tremendous value of his long runs. Sanders is a great example.

Let's make a little experiment. Suppose we take any AP game, say the last Viking loss, in the first Packer game, in which AP had 21 carries for 210 yards. If I told you you could distribute his yardage among his 21 carries any way you wanted, how would you do it? For example, you might give him 5 yards in each of his first two carries to help the Vikings avoid a three-and-out in their first offensive series.

Meanwhile, look at the 2nd offensive series. AP had 3 yards on first down, 5 yards on 2nd and 7, 8 yards on 3rd and 2 (on 3 consecutive carries), 3 yards on first down, 3 yards on first down, 2 yards on 3rd and 1, 9 yards on 1st and 10. That's a total of 7 carries for 33 yards and no big plays (longest play of the drive was a 13-yard completion by Ponder). But AP had two 3rd down conversions, no stuffs, and about half the drive yardage. The result of the drive? TD Vikings.

What if he had been stuffed on those 3rd down plays, even if there had been a long run involved earlier in the drive? They'd have a FG, at best.

------
FO posters are a peacock. You got to let us fly!

9
by Vincent Verhei :: Tue, 12/25/2012 - 1:17am

Maybe I missed this over the years, but is there a reason why FO doesn't choose to report median YPC instead of mean YPC? It would help encapsulate both average yardage and success rates while reducing the influence of long carries and probably better correlate to DVOA and DYAR.

Another reader addressed this, but just for fun I checked the median carry for the top five guys in total carries:

Arian Foster: 3
Adrian Peterson: 3
Alfred Morris: 3
Marshawn Lynch: 3
Doug Martin: 3

Well, the specific number there is a yard higher than expected, but otherwise the result is the same. For comparison, the lowest-ranked regular RB by DYAR is Darren McFadden. His median YPC? 2. So that appears to be the entire range of possibilities, and it's not very useful.

I did experiment one time with dividing each RB's carries into three equal groups — long, medium, and short — and calculating the average yards per carry of each group. Maybe I'll revisit that after the season.

How did the Chiefs RB's rank before defensive adjustments? Was it the Colt's defensive modifier that dragged them out of the top 5?

Without opponent adjustments, Hillis would have been eighth, Charles tenth. Hillis had four runs for 10 yards or more, seven total first downs, and only two stuffs. Charles had four 20-yard runs, bu only one other first down, and he had a fumble. Most importantly, both players had negative receiving DYAR, with a combined one catch for 4 yards in four targets. In rushing value only, and without opponent adjustments, they would have ranked fifth and sixth, with Charles ahead of Hillis by just a few decimal points.

Absolutely correct. Kaepernick was about to throw it and pumped, I think because Aldon Smith was jumping into the air to get a hand on the screen pass. Kaepernick paused and pulled down the ball, and leaned back to throw around Smith. As a result, two defenders were waiting to plaster Manningham. Kaepernick either needed to lob over Smith, get the pass off before Smith was prepared to jump, or bail out and throw somewhere else.

Forgive me for being a jerk, but I'm pretty certain Aldon Smith wasn't trying to knock Kaepernick's passes down, since they're on the same team and everything.

Every week you have to write why a QB who played a terrible or below average game is near the top because they played a good defense.

It's not that they're near the top because they played a great defense, it's that they're higher than their raw numbers would suggest because they played a good defense. It's especially important to mention this for quarterbacks who play the Bears. Chicago has come back to the pack somewhat, but earlier in the year they were skewing things so badly that there were only six or seven defenses in the entire league that graded out as "above average."

16
by Arkaein :: Tue, 12/25/2012 - 3:18pm

I'm not at all surprised that career or season median RB gains are not very interesting, but I'll bet that per-game numbers would be quite illuminating, with high medians likely well correlated to high success rates.

This would be true for passing plays as well as rushing. Other stats that might be similarly interesting would be stats like the mean of the middle, say, 80% of runs (throw out the high and low 10%), which would still give results with finer granularity than full yards.

22
by Sifter :: Wed, 12/26/2012 - 1:33am

Yes, per game median would be much more interesting. Medians work better with a smaller sample size ie. 16 games rather than 300 carries. That's why I'm a big proponent of having an extra DVOA table added - the same DVOA numbers, but calculated in medians instead of means. Would help dilute big outlier games, would make the future schedule number make more sense and just be a useful addition in general I feel. The schedule thing bugs me ie. 3 very hard games, and 7 pretty easy ones might equal a middling schedule on the DVOA table, when in reality it's a pretty easy schedule.

11
by BlueStarDude :: Tue, 12/25/2012 - 10:32am

The big difference between Peterson and the names at the top of the rushing DYAR list: all those top performers played on teams with potent passing attacks. DYAR is underestimating Peterson's performance this year because it can't account for the fact that (without Harvin for most of the season) he's the only real big play threat on his offense.

12
by The Hypno-Toad :: Tue, 12/25/2012 - 10:43am

Probably some truth to that.

14
by coremill (not verified) :: Tue, 12/25/2012 - 12:12pm

Terrell Davis, Smith, Faulk, Holmes, and Johnson also all ran behind at least very good if not all-time great lines that featured multiple pro bowlers/all-pros/potential Hall of Famers. It is not a coincidence that the mid-90s cowboys, late 90s Broncos, and early-00s Chiefs show up multiple times on this list -- those were some of the best run-blocking lines of all time.

The Minnesota line this year may grade out decently but not one of its members has ever been to a pro bowl or made all-pro and they are nowhere near the level that any of those lines were.

19
by The Ninjalectual :: Tue, 12/25/2012 - 4:24pm

When your left tackle is a rookie, not having been to a pro-bowl doesn't mean much.

13
by Nathan :: Tue, 12/25/2012 - 11:50am

Brian Hoyer! How did I miss that? The things you don't see when you're home for Xmas and don't have NFL ticket.

17
by IHeartAdderall (not verified) :: Tue, 12/25/2012 - 3:22pm

Yep, I had no idea he was in AZ. I still remember, a year or two ago, when Simmons had Mike Lombardi on his podcast, and there was supposedly some talk of Lombardi being named GM in San Francisco (so, two years ago, I guess), and Lombardi said one of the first things he'd do if he got the job was to trade for Brian Hoyer. I assumed it was because Lombardi/Simmons always stain their jeans when discussing the Patriots, but thought there could be something to it. The fact that, now, Hoyer is a late-season signing by AZ kinda tells me there wasn't. It also tells me that SF dodged a bullet by not hiring Lombardi.

18
by dmstorm22 :: Tue, 12/25/2012 - 3:43pm

Apparently Lombardi/McDaniels is rumored to be a possible package to Cleveland, and their first move might be trading for Ryan Mallett.

I like the Browns. I wish them well. I want them to do well, but there is no way that trinity works out in any way.

25
by little bobby tables (not verified) :: Wed, 12/26/2012 - 3:56pm

to that, I say a hearty LOLOLOLOL. the patriots are the rays of the NFL: If they agree to a trade with you, don't do it. You'll come out on the losing end every single time.

McDaniels is an interesting and innovative coach, but he needs to be paired with a strong-handed GM. That is not a description that one would attach to rookie GM Mike Lombardi.

26
by Nathan :: Wed, 12/26/2012 - 4:58pm

I think I'm higher on Mallett than most Pats fans, but that would bum me out, I've liked what I've seen in the preseason and I want to see them groom him long term, Rodgers style.