Pollard, Perine, and the NFL's Top Receiving Backs

Dallas Cowboys RB Tony Pollard
Dallas Cowboys RB Tony Pollard
Photo: USA Today Sports Images

NFL Week 11 - The statistical stars of Week 11 were a pair of running backs who didn't do a ton of running—at least, not as running is defined in the NFL play-by-play. But Tony Pollard and Samaje Perine were so deadly as receiving threats that they carried their teams to victory anyway.

Pollard finishes as the best player at his position in our rankings this week, one of the many Cowboys who shined in Dallas' 40-3 demolition of the Minnesota Vikings. His surface-level rushing numbers are fine enough (15 carries for 80 yards, a 5.3-yard average), but while he did rush for five first downs, he was more explosive on the ground than reliable. His totals are boosted by three long runs that gained 17, 18, and 20 yards, but he was also stuffed for no gain or a loss three times, and half of his carries gained 3 yards or less. That all adds up to 13 rushing DYAR, not even among the top 10 running backs this week.

When the Cowboys threw Pollard the ball, however, things went much better. He caught each of the six targets thrown his way, and while they didn't always work out—one catch went for no gain on second-and-3, and another lost 3 yards on second-and-11—he made enough big plays to make it work. Four of his catches produced first downs, including a 30-yard touchdown on second-and-7 and a 68-yard score on third-and-14. Five of those catches came on screen-type receptions behind the line of scrimmage, but the longest was caught downfield, with 22 yards through the air and 46 coming after the catch. The happy totals: six targets, six catches, 109 yards, two touchdowns, 63 receiving DYAR. That latter number is the best for any running back in a game this season.

The season's second-best receiving game by a running back was actually going on at the exact same time in Pittsburgh. While Pollard was boom-or-bust against the Vikings, Perine was mostly bust on the ground in the Bengals' 37-30 win over the Steelers—11 carries, 30 yards, a 2.7-yard average, a long run of just 8 yards, only two first downs, a pair of stuffs, finishing below replacement level in rushing DYAR. But Cincinnati also threw him the ball four times. The results?

  • 29-yard touchdown on second-and-9.
  • 6-yard gain on first-and-10.
  • 11-yard touchdown on first-and-10.
  • 6-yard touchdown on second-and-goal.

Each of those touchdowns was caught behind the line of scrimmage, with Perine putting in most of the work after the catch. The first two broke ties on the scoreboard to put the Bengals ahead, while the third extended a fourth-quarter lead from four points to 11, effectively icing a Cincinnati victory. Perine's totals: four targets, four catches, 52 yards, three touchdowns, 61 receiving DYAR.

Pollard and Perine are the first running backs this year to top 60 receiving DYAR in a game. Four players did it last year, led by Cordarrelle Patterson's three-touchdown game against Washington in Week 4, but that was an anomaly, not part of a trend. There were three such games in 2019, but none in 2018 or 2020. Usually there are one or two in each season, with 10 in the 1980s (a decade with two strike years, as well as a 1980 campaign that we haven't processed yet), 16 in the 1990s, and 16 again in the 2000s, but only 12 in the 2010s. Marshall Faulk, of course, leads all running backs with six, twice as many as anyone else.

But that's not to say that what Pollard and Perine did was common. Per Stathead, Pollard is the first running back to catch 100% of his passes for at least 100 yards and two scores in nearly a decade, since Jamaal Charles' eight-catch, 195-yard, four-touchdown game against the Oakland Raiders in Week 15 of the 2013 season. (With end-of-season adjustments, Charles had 112 receiving DYAR that game, setting a record for running backs that still stands.) Perine, meanwhile, is the first back with a perfect catch rate for at least 50 yards and three scores since Danny Woodhead in 2015. So no, statlines like this are not typical.

What's interesting here is that neither Pollard nor Perine are technically first-string players. Pollard has only started three games this season, but he has split time pretty evenly with Ezekiel Elliott, each playing roughly 35 snaps on offense per game. Elliott gets more carries, which frustrates some Cowboys fans because Pollard has been so much more productive, but Pollard has caught three times as many passes and ranks seventh or higher at his position in both DYAR and DVOA as a receiver.

In Cincinnati, Perine is the clear backup behind Joe Mixon, who has doubled Perine in snaps and targets this year and quadrupled him in carries. (Mixon left Sunday's game with a concussion, and Perine saw a season-high 46 offensive snaps as a result.) With only 26 targets, Perine barely qualifies for our running back receiving leaderboard—but qualify he does, and among those 38 qualifiers he ranks fourth in DYAR and first in DVOA. That can happen when you catch 84.6% of your passes for 6.8 yards per throw with four scores. Mixon, by the way, has also been an effective receiver, ranking sixth in DYAR, and Joe Burrow ranks second among quarterbacks in passing DYAR on throws to running backs (foreshadowing!).

The only passer who has done more damage than Burrow on throws to runners this season is Jimmy Garoppolo, who has gotten to play with the best receiving back in the NFL, at least according to DYAR. After Monday Night Football, Christian McCaffrey leads all backs with 141 DYAR as a receiver, leading all players at the position with 462 yards and ranking second with 67 targets and 54 catches. (Austin Ekeler of the L.A. Chargers is first in those categories with 83 targets and 69 catches, for 437 yards, but he is second in DYAR because he is averaging 2 full fewer yards per catch than McCaffrey and also has a fumble.) McCaffrey is catching about one less ball per game with the 49ers than he did with the Panthers, but his catch rate has climbed from 76.7% to 87.5%, and his yards per catch has jumped from 8.4 to 8.8. McCaffrey had 59 DYAR in six games with Carolina, but 82 DYAR in only four games with San Francisco. Other 49ers running backs who have been effective receivers in small sample sizes this year include Kyle Juszczyk (15 targets, 11 catches, seven first downs, 13.9 yards per catch) and Tevin Coleman (three catches—all of them against McCaffrey and the Panthers, coincidentally—for 44 yards and a touchdown).

McCaffrey's opposite has been Indianapolis' Jonathan Taylor, last by a mile in both DYAR (where no qualifying back has been even half as bad) and DVOA. And while the entire Indanapolis offense has been a disaster this season, this appears to be a Taylor problem more than a Colts problem. Nyheim Hines (before he was traded to Buffalo) and Deon Jackson both have positive DYAR on 20-plus targets for the team this season. But Taylor has seen 31 targets (only one—an incompletion on third-and-2—from Sam Ehlinger) and turned them into 21 catches for 97 yards, a pathetic average of only 4.6 yards per reception. Only three of those catches gained first downs, and only one gained 10 yards or more. Despite missing two games, Taylor is on pace to hit negative triple digits in receiving DYAR. Running backs have only hit that mark 11 times on record, the last being Ben Tate with the Houston Texans in 2013. Things haven't significantly improved for Taylor under Jeff Saturday, either—in his last two games, he has caught five of six passes for only 26 yards. That includes a three-catch, 10-yard outing against Philadelphia that made him the least valuable running back this week. Let the watch for -100 begin!

Quarterbacks
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
1.
Jacoby Brissett CLE
28/41
324
3
0
1
242
256
-14
BUF
There are a few reasons Brissett makes a surprise appearance in first place this week. First, opponent adjustments: he was one of four quarterbacks to gain at least 50 DYAR based on who they played. Second, garbage time: he had the best passing DYAR in the fourth quarter, and after the Bills took a 28-10 lead with less than seven minutes left in the game, Brissett went 10-of-14 for 121 yards and two touchdowns. Now, to be fair to Brissett, we should point out the Cleveland jumped out to a 10-3 lead in part because he also had the best passing DYAR in the first quarter, when he went 8-of-9 for 102 yards and a touchdown.
2.
Derek Carr LV
23/37
307
2
0
1
165
165
0
DEN
Things were not easy for Derek Carr. He had to play the stout Broncos defense, gaining 56 DYAR due to opponent adjustments, second-most this week. He was often in long yardage, with a league-high average of 11.5 yards to go for a first down. Despite those handicaps, he still finished as the week's best passer from under center (9-of-14 for 149 yards and two touchdowns) and on deep balls (6-of-9 for 171 yards and two scores).
3.
Dak Prescott DAL
22/25
276
2
0
0
146
145
1
MIN
Prescott averaged 11.0 yards per dropback and was successful 72% of the time, best in the NFL in both categories this week. His receivers certainly helped—his average completion gained a league-best 7.2 yards after the catch. He was at his best in the second quarter, going 9-of-10 for 105 yards and a touchdown.
4.
Jimmy Garoppolo SF
20/29
228
4
0
0
142
140
2
ARI
5.
Ryan Tannehill TEN
22/27
333
2
1
3
133
133
0
GB
Tannehill's average dropback came with only 7.2 yards to go for a first down, least in the league among qualifying passers. (If you include non-qualifiers, Tannehill falls to second behind his teammate, Derrick Henry.) He was nearly perfect on throws to his left, going 13-of-14 for 208 yards and a touchdown.
6.
Patrick Mahomes KC
20/34
326
3
0
1
125
142
-17
LAC
Mahomes led all quarterbacks in DYAR on third/fourth downs (5-of-9 for 115 yards and five conversions, including a 32-yard touchdown) and on throws to tight ends (10-of-15 for 181 yards and three touchdowns).
7.
Andy Dalton NO
21/25
260
3
0
3
118
118
0
LAR
Perfection, thy name be Andy Dalton, at least when it comes to throws down the middle for one afternoon against the Los Angeles Rams. Dalton threw five passes down the middle of the L.A. defense, completing each of them for a first down and a total of 101 yards, including 7- and 53-yard touchdowns. He was also the week's best passer in the third quarter, when he went 8-of-9 for 145 yards with two touchdowns and one sack.
8.
Joe Burrow CIN
24/39
355
4
2
2
105
102
3
PIT
As foreshadowed earlier, Burrow was the top passer this week on throws to running backs, completing seven of his eight passes for 94 yards and three touchdowns. As not foreshadowed, he was also the top passer in the red zone, with three dropbacks resulting in three completions for three touchdowns, for a total of 18 yards.
9.
Aaron Rodgers GB
25/39
227
2
0
1
59
59
0
TEN
Rodgers' last pass of the third quarter was an 8-yard touchdown to Christian Watson that cut the Tennessee lead to 20-17. By the time he got the ball back, however, the Titans had stretched that lead out to 27-17, and Rodgers and the Packers never really challenged them again. In the fourth quarter, he went 8-of-16 for only 60 yards with a sack.
10.
Matthew Stafford LAR
11/18
159
2
0
1
56
59
-3
NO
Stafford's average pass traveled an NFL-high 12.1 yards downfield. Before he left the game midway through the third quarter, he threw five deep balls, completing two of them for 82 yards and a touchdown.
11.
Josh Allen BUF
18/27
197
1
0
2
56
58
-3
CLE
Allen had a rough day on third downs, going 2-of-7 for 17 yards with more sacks (two) than conversions (one).
12.
Justin Herbert LAC
23/30
280
2
1
5
54
50
3
KC
For most of this game, Herbert was throwing to the outside. The few times he did throw down the middle, he didn't have much success, going 3-of-5 for only 20 yards.
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
13.
Jalen Hurts PHI
18/25
190
1
0
3
51
7
44
IND
Hurts leads all quarterbacks in rushing DYAR after running 13 times for 88 yards and eight first downs, including a 7-yard game-winning touchdown on third-and-goal. He wasn't nearly as effective as a passer, however. In the third quarter, in fact, he was the league's worst passer—though he completed 5-of-6 passes for 43 yards, only one of those completions picked up a first down, and he also gave up two sacks and two fumbles.
14.
Mac Jones NE
23/27
246
0
0
6
50
51
0
NYJ
Jones was another quarterback who struggled on third downs. He completed six of his eight passes for 66 yards, but only two of those completions moved the chains, and he also gave up a pair of sacks.
15.
Marcus Mariota ATL
13/18
131
1
0
0
49
52
-3
CHI
What a streaky day this was. Mariota started 5-of-6 for 49 yards and a touchdown. Then he completed one of his next four passes for only 2 yards. Then he completed seven passes in a row for 80 more yards before his final pass fell incomplete.
16.
Matt Ryan IND
23/32
213
0
0
4
46
46
0
PHI
Ryan gains a league-high 59 DYAR due to opponent adjustments. He threw for only one first down inside the Philadelphia 40, going 3-of-6 for 46 yards with two sacks.
17.
Jared Goff DET
17/26
165
0
0
0
46
43
3
NYG
Goff missed only one pass when throwing to his right, going 8-of-9 for 74 yards.
18.
Kenny Pickett PIT
25/42
265
1
0
2
39
41
-3
CIN
Pickett threw 12 failed completions, tied for most in the NFL this week. Three of them came inside his own 20-yard line, where he went 3-of-4 for only 8 yards.
19.
Russell Wilson DEN
24/31
247
0
0
3
15
16
-1
LV
Wilson loses a league-high 69 DYAR due to opponent adjustments. He was in the top 10 in passing DYAR in the first half (13-of-15 for 147 yards, plus an 18-yard DPI), but third-worst after halftime (11-of-16 for 100 yards, plus three sacks).
20.
Taylor Heinicke WAS
15/27
191
0
0
0
11
41
-30
HOU
Heinicke's three runs netted a loss of 7 yards, including a fumble. The Commanders still won this game by 13 points, and could have won by more if they didn't have to settle for three field goals. That's partly on Heinicke, who only completed one pass for 5 yards in five red zone throws.
21.
Daniel Jones NYG
27/44
341
1
2
2
-10
-27
17
DET
Jones loses 58 DYAR due to opponent adjustments. He ran seven times against the Lions, picking up 50 yards and five first downs, including a 3-yard touchdown that gave New York a lead in the first quarter. Unfortunately that lead didn't last long, in part because Jones was the week's worst passer in the second quarter, going 6-of-14 for 84 yards with an interception and a sack.
22.
Trace McSorley ARI
6/10
59
0
1
0
-21
-25
4
SF
All of McSorley's action came with Arizona down 38-10 in the fourth quarter.
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
23.
Bryce Perkins LAR
5/9
64
0
0
3
-25
-34
8
NO
Perkins came into the game with L.A. down 24-14 in the third quarter. He promptly ripped off a 23-yard run on second-and-10, gaining more rushing yards on one play than Matthew Stafford has had all season, and went on to add 16 more yards on four more carries. As a passer, he mixed in some chunk plays with some sacks, and offered no kind of threat once crossing midfield—he only completed one of his five passes in Saints territory, and that one was an 8-yard gain on second-and-18.
24.
Colt McCoy ARI
24/34
218
0
1
3
-37
-40
2
SF
25.
Justin Fields CHI
14/21
153
1
1
4
-45
-40
-6
ATL
Fields loses 40 DYAR due to opponent adjustments. He ran for 81 yards, but it took him 18 carries (!!!) to get there, and he also had a fumble on a running play. He was the league's worst passer on throws down the middle, going 2-of-5 for 19 yards with more interceptions (one) than first downs (zero).
26.
Zach Wilson NYJ
9/22
77
0
0
4
-46
-58
12
NE
Wilson averaged 1.7 yards per dropback and he was successful only 19% of the time, worst in the NFL in both categories this week. On first and second downs, he went 4-of-13 for 19 yards with more sacks (one) than first downs (zero). He only threw for three first downs in the entire game, and all three were third-down conversions, but then he was also sacked three times on third down. Two of those first downs came on consecutive throws in the second quarter; after that, he went 6-of-14 for only 24 yards with four sacks. He threw only one pass in New England territory: an incompletion on third-and-2. He threw six passes to receivers at or behind the line of scrimmage, completing only three of them, and those three gained only 5 yards. You get the point. And yes, because Wilson only had 26 dropbacks, there were four quarterbacks worse than this in Week 11.
27.
Lamar Jackson BAL
24/33
209
0
1
3
-56
-62
6
CAR
Several times this year, Jackson has finished as the NFL's top passer on throws to tight ends. Well this week he finished last in that category, going 8-of-12 for 66 yards with an interception. He wasn't much better on throws to running backs, going 6-of-8 for only 12 yards. Fortunately for the Ravens, passes to wide receivers count too, and Demarcus Robinson … well, we're getting ahead of ourselves.
28.
Kirk Cousins MIN
12/23
105
0
0
7
-66
-66
0
DAL
It could have been worse for Cousins—the Vikings mercifully pulled him from the game after he took his seventh sack late in the third quarter. He spent a lot of the afternoon dumping off to checkdown options who weren't really open; his average completion gained a league-low 2.4 yards after the catch. He had a terrible day on third/fourth downs, going 2-of-5 for 20 yards with four sacks, one fumble, and only one conversion. In his defense, Cousins is not the first quarterback to struggle against Dallas this year—he gains 53 DYAR due to opponent adjustments.
29.
Baker Mayfield CAR
21/33
196
0
2
4
-132
-135
3
BAL
And now, a series of related statistics:
  • Mayfield was the week's worst passer on throws to running backs, going 4-of-7 for 39 yards with an interception.
  • Mayfield was the week's worst passer on throws to receivers at or behind the line of scrimmage, going 6-of-11 for 28 yards with a pair of interceptions.
  • Mayfield threw 12 failed completions, tied for most in the league this week.
  • Four of those failed completions came on third/fourth downs, where Mayfield also finished last in DYAR, going 5-of-8 for 54 yards with only one conversion, plus a 12-yard DPI, an interception, and two sacks.
30.
Davis Mills HOU
19/33
169
0
2
5
-149
-170
22
WAS
Mills was much better as a runner (four carries, 11 yards, one touchdown) than he was as a passer, but rarely has anyone been damned with praise so faint. Mills didn't throw for a first down until the Texans were down 20-0 in the third quarter; by that point, he had gone 7-of-10 for 31 yards while giving up 23 yards on three sacks and also throwing a pick-six. He finished as the league's worst passer on throws to his left (8-of-14 for 92 yards with an interception) and throws to his right (6-of-12 for 30 yards with an interception). He was worst on deep balls (0-for-5 with an interception) and also worst in the red zone, where it's almost impossible to throw deep (2-of-7 for 16 yards, no touchdowns, one sack). A bad, bad, bad, bad day.
Five Best Running Backs by DYAR (Total)
Rk
Player
Team
Runs
Rush
Yds
Rush
TD
Rec
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
Opp
1.
Tony Pollard DAL
15
80
0
6/6
109
2
75
12
63
MIN
2.
Samaje Perine CIN
11
30
0
4/4
52
3
59
-1
61
PIT
3.
Josh Jacobs LV
24
109
0
3/4
51
0
46
23
23
DEN
Although the Broncos stuffed Jacobs five times, they also allowed him to run for eight first downs, three of which gained at least 10 yards. Only one of his three catches produced a first down, but that one was a 43-yard gain with Las Vegas down by three points and less than a minute to go.
4.
Najee Harris PIT
20
90
2
4/6
26
0
35
26
9
CIN
Harris only ran for three first downs against the Bengals, but two of those were touchdowns of 1 and 19 yards, and the other was a 13-yard gain on first-and-10. Further, Cincinnati only stuffed him twice, and nine of his runs counted as successful plays. His best catch was an 11-yard gain on third-and-7.
5.
Damien Harris NE
8
65
0
2/2
28
0
34
17
17
NYJ
The fifth-best running back (and second-best Harris) of Week 5, Damien was stuffed twice by the Jets while only running for two first downs, but those two first downs gained 30 and 22 yards, both on second-and-5. Both of his catches—13- and 15-yard gains on first-and-10—also moved the chains.
Five Best Running Backs by DYAR (Rushing)
Rk
Player
Team
Runs
Rush
Yds
Rush
TD
Rec
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
Opp
1.
Najee Harris PIT
20
90
2
4/6
26
0
35
26
9
CIN
2.
Josh Jacobs LV
24
109
0
3/4
51
0
46
23
23
DEN
3.
Dalvin Cook MIN
11
72
0
0/1
0
0
17
23
-6
DAL
All of Cook's carries gained at least 2 yards. Three of them gained first downs on gains of 11, 11, and 17 yards.
4.
Kareem Hunt CLE
5
32
0
2/2
22
0
32
22
10
BUF
Each of Hunt's five carries gained at least 3 yards and counted as a successful play. Three went for first downs, the longest a gain of 11.
5.
Elijah Mitchell SF
9
59
0
0/0
0
0
20
20
0
ARI
Worst Running Back by DYAR (Total)
Rk
Player
Team
Runs
Rush
Yds
Rush
TD
Rec
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
Opp
1.
Jonathan Taylor IND
22
84
1
3/4
10
0
-57
-44
-13
PHI
We mentioned Taylor's struggles as a receiver earlier, but let's not overlook how ineffective he was on the ground at the same time. Only three of his 22 carries resulted in first downs, none of which came in the second half. He would have had a fourth, but he fumbled after a 7-yard gain on third-and-1. Meanwhile, he was stuffed seven times. Finally, he loses 17 DYAR for playing the Eagles and their previously porous run defense … but if Linval Joseph and Ndamukong Suh keep playing like this, those opponent adjustments won't be so strong by Week 18.
Worst Running Back by DYAR (Rushing)
Rk
Player
Team
Runs
Rush
Yds
Rush
TD
Rec
Rec
Yds
Rec
TD
Total
DYAR
Rush
DYAR
Rec
DYAR
Opp
1.
Saquon Barkley NYG
15
22
0
2/5
13
0
-55
-52
-3
DET
This is the second week in a row that Barkley has been the NFL's least valuable rusher, as opponents have figured out that if you stop the ground game, you stop the Giants. He failed to run for a single first down against the Lions. His only successful carry, a 4-yard gain on second-and-4 from the 7, was also tied for his longest run of the day. Meanwhile, he was stuffed five times.
Five Best Wide Receivers and Tight Ends by DYAR
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
Opp
1.
Demarcus Robinson BAL
9
9
128
14.2
0
59
CAR
Six of Robinson's catches picked up first downs and three gained 20 yards or more, the longest a gain of 31. The three that didn't pick up first downs? They were 6- and 7-yard gains (which still count as successful plays) and a 9-yard gain third-and-13 (which doesn't, but is practically neutral for DYAR purposes).
2.
Amari Cooper CLE
8
12
113
14.1
2
53
BUF
Seven of Cooper's catches produced first downs, including 7- and 25-yard touchdowns and a pair of third-down conversions.
3.
Davante Adams LV
7
13
141
20.1
2
52
DEN
Six of Adams' catches produced first downs and four gained 20 yards or more, including a 31-yard touchdown in the second quarter and the 35-yard game-winner in overtime.
4.
Donovan Peoples-Jones CLE
5
6
61
12.2
1
50
BUF
All five of Peoples-Jones' catches counted as successful plays, and three of them—20- and 27-yard gains, plus a 2-yard touchdown—picked up first downs. He added a fourth first down on a 36-yard DPI.
5.
Tee Higgins CIN
9
13
148
16.4
0
50
PIT
Eight of Higgins' nine catches produced first downs; the other was a 6-yard gain on first-and-10. Those eight catches that did move the sticks each gained at least 12 yards, the longest a 33-yard gain on second-and-2 in the third quarter.
Worst Wide Receiver or Tight End by DYAR
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
Opp
1.
Laviska Shenault CAR
4
4
7
1.8
0
-26
BAL
It's not often that the worst receiver of the week has a 100% catch rate with no fumbles, but none of Shenault's four receptions (1-yard loss on first-and-10, 5-yard gain on third-and-10, 1-yard gain on first-and-10, and 2-yard gain on second-and-10) picked up first downs or even counted as successful plays.

Comments

22 comments, Last at 23 Nov 2022, 12:23pm

#1 by HitchikersPie // Nov 22, 2022 - 7:56am

Beyond the obvious disgusting behaviour from Watson in the past, he's been out of football for a year and a half, and Brissett is playing some of the best football of his life, with only his second positive DVOA season, and 2.6% in 2019 hardly compares to the 17.1% he's putting up this year. Naturally it follows his passing DYAR is a massive career high, but even his rushing DYAR is higher than ever, 76 so far vs a prior high of 57. PFF agrees this is his best season, and first with a grade over 80 (roughly pro bowl level), and again by EPA he's already doubling up his best EPA/play and almost 15 more EPA through half a season than his previous seasons best.

 

I'm sure the Browns will go to Deshaun, but it won't be too surprising if their multimillion-dollar QB is less impressive than JaGOATby.

Points: 0

#2 by Mike B. In Va // Nov 22, 2022 - 8:20am

I was impressed by some of his throws in this game - on time and on the money, and good decision-making. This was my first time seeing him this year, and I don't think he's the reason they're 3-7.

Points: 0

#3 by Lost Ti-Cats Fan // Nov 22, 2022 - 8:45am

I wonder who's more likely to carry his 2022 performance into 2023, Geno Smith or Jacoby Brissett?

Points: 0

#4 by af16 // Nov 22, 2022 - 8:55am

What was Lamar’s opponent adjustment?

Points: 0

#5 by Pat // Nov 22, 2022 - 9:54am

Second, garbage time: he had the best passing DYAR in the fourth quarter, and after the Bills took a 28-10 lead with less than seven minutes left in the game, Brissett went 10-of-14 for 121 yards and two touchdowns.

I know plenty of others don't agree, but to me it's a stretch to call that "garbage time." I think you have to go way into the "this game is totally gone" range to get to what should be called garbage time - like "pull the starting QB on the losing team" range, a la Vikings-Cowboys. That situation really rarely happens in the NFL (although it's obviously ultra-common in college).

The Browns-Bills late game just felt like standard "prevent on defense, burn clock on offense," and I don't think it makes sense to dismiss Brissett's performance there because it forced Buffalo to actually need to do something on offense. Allen dropped back to pass twice on the final drive, and got sacked on one of them.

I don't want to say it "clearly" feels like it wasn't garbage time - it feels like "just on the border of garbage time." The last drive I think is a tough situation because it feels like it's "neutral" QB performance - not bad, but not great or anything, and I'm not necessarily sure that giving a boost for playing a great defense in that situation makes sense.

Points: 0

#6 by Mike B. In Va // Nov 22, 2022 - 10:15am

Allen dropped back to pass twice on the final drive, and got sacked on one of them.

Took a sack to keep the clock running. He could have thrown it away easily.

Points: 0

#16 by Pat // Nov 22, 2022 - 9:39pm

I was trying to point out the possible risk, not the actual one.

Points: 0

#7 by Aaron Brooks G… // Nov 22, 2022 - 10:17am

The Browns-Bills late game just felt like standard "prevent on defense, burn clock on offense," 

This is basically when the game state starts to violate DVOA's assumptions about how the game works.

Once you enter an end-game state where you are trading time for yardage, DVOA's in-game assessment begins to break down, because DVOA assumes yardage represents the game state. I know the aggregate predictiveness argument, but if we're discussing what is and is not garbage-time, strategic time-wasting is a good criterion, because that's when WinPct considerations trump EPA and yardage.

Points: 0

#10 by Oncorhynchus // Nov 22, 2022 - 12:49pm

I don't know that this disconnect of time vs yardage in EPA-based metrics is limited to just the late game. Consider the Eagles/Commanders game from Week 10. The Commanders pursued a strategy heavily dependent on the run game for two reasons: 1) the Eagles had a weak run-defense they could exploit, and 2) they wanted to deny the Eagle's offense possessions. This was a defensive strategy on offense that worked. And I think it works against other prolific passing offenses like the Chiefs and the Bills. The caveat is that obviously this strategy only works in a near-neutral situation (tie game) or when the time-wasting offense is leading.

I think the second point of the Commanders strategy is underrated in analytics. That's why I wonder what the value of a possession is in EP-models. It's in there if you look at the play-by-play data. A drive that starts at 25 has higher EP at first snap of the drive than a drive that starts at the 10. I think there's also time adjustments in some EP-models. A drive that starts at the 25 should have a lower EP if there are 30 seconds remaining in the 2nd quarter versus 3 minutes. I know people make a big-deal out of the fact that throwing the ball almost always has a higher EPA/play than running the ball. But what bothers me about that analysis (from a statistics perspective) is I think it ignore the effect that a run-heavy game has in denying the opponent possessions.

Consider two hypothetical 2nd quarter drives in a tie-game. Both start at the 25 after a touchback. One drive results in a touchdown after 3 huge passing plays each, the other drive results in a touchdown after 15 plays with more of them being runs than passes. The first drive leaves 3 minutes on the clock, the second drive leaves 30 seconds. DVOA and EPA/play would credit the team who had the first drive as being more efficient. They're the "better" offense. But what if you were to include the expected points you left your opponent?

To break it down further: let's say in a neutral situation a drive starting at the 25 with plenty of time left on the clock is worth 1.5 expected points. So in our two hypothetical drives, both added 5.5 points over expectation. EPA/drive is 5.5 if we ignore the fact that you have to give your opponent the ball back after you score. But let's say a drive starting at the 25 with 30 seconds left before the end of half or end of game is worth only 0.5 expected points. If you were to credit that reduction in EP to the offense (as you should) the second drive is more valuable on a per possession basis, but not on a per play basis. But the nerds (relax, I am one) would tell you that the first drive was a sign of a clearly better offense. I think that's wrong. And it's wrong from an analytics perspective, not a "do you even watch tape, bro?" perspective.  

Now obviously you'd want to adjust for opponents in any realistic model. A Chiefs drive starting at the 25 has more expected points than a Broncos drive starting at the 25. I think a good model would account for the fact that an offense that denies the Chiefs possession in neutral or leading situations is better than an offense that merely goes toe-to-toe with the Chiefs offense. Especially if said offense can do both. I'd argue that the best offensive teams are those that have a high EPA/play when trailing, but a high opponent-possession-adjusted EPA/drive when leading or tied.

I'm not exactly sure how you'd model this, but you'd probably want to include some running tally of expected-possessions. This would change a lot of metrics. Running the ball probably looks more valuable if you modeled it as expected-points-added + expected-opponent-possessions-denied (this is especially true if pass-play have a higher turnover probability; I'm not sure if they do). A quick 3-and-out would have an even more negative EPA because not only do you remove the expected-point the model gave you at the start of the drive, but you might actually increase the expected-possessions of your opponent.

(Also I don't know that WinPct models ae a good adjustment for accounting for time of possession. Because it'll barely budge in the 1st half.)

Points: 0

#11 by Aaron Brooks G… // Nov 22, 2022 - 1:09pm

Consider the Eagles/Commanders game from Week 10. The Commanders pursued a strategy heavily dependent on the run game for two reasons: 1) the Eagles had a weak run-defense they could exploit, and 2) they wanted to deny the Eagle's offense possessions. This was a defensive strategy on offense that worked. And I think it works against other prolific passing offenses like the Chiefs and the Bills. The caveat is that obviously this strategy only works in a near-neutral situation (tie game) or when the time-wasting offense is leading.

I would argue that running teams often abandon the run too early.

When the Eagles went down 9 to the Redskins late in the 3rd, they started running the ball. They ran 13 times to 4 passes before the Goedert facemask, and then the one successful play when down 5 was also a rush. 

Against Indy, starting when they went down 10, they ran 20 times -- 11 consecutive rushes in one stretch. You can dig out of a big hole even by continuing the run the ball, so long as your possessions are fruitful.

It's interesting that running is considered a Goliath Strategy. Get a lead, then sit on it by rushing the clock out. At one time, it probably was. Now, I think it's a viable David Strategy.

David Strategies are usually considered to be those that increase variance. This suggests that you should start mad bombing passes -- the ultimate boom/bust strategy. But you can also increase variance by decreasing sample size. The Princeton Offense in basketball functions as a high-variance David Strategy in basketball -- it has a really low EPA/second, but a really high EPA/possession. This is what Washington was doing. In terms of points/play or points/minute, the Eagles smoked them. In terms of points/possession, Washington had a big advantage. If you can keep being efficient on a per-drive basis, shortening the game is a really effective David Strategy, because Goliath can't use his advantages to run the score away from you.

Points: 0

#22 by turbohappy // Nov 23, 2022 - 12:23pm

Even if you aren't so efficient per drive shortening the game can be a good David strategy. The less drives, the more effect random chance has. A couple fumbles or a muffed punt or something and you can win a game as a dramatically less efficient team whereas if both teams get more drives you lose every time.

Points: 0

#12 by Aaron Brooks G… // Nov 22, 2022 - 1:11pm

(this is especially true if pass-play have a higher turnover probability; I'm not sure if they do)

They should. Interceptions outnumber fumbles lost, and you figure a substantial fraction of lost fumbles also occur on passing plays.

Points: 0

#15 by Pat // Nov 22, 2022 - 9:38pm

I disagree on this, because prevent *is* normal game state there and VOA's only comparing to other prevent situations. There's plenty of examples of prevent, it's easy to know what "normal" vs prevent looks like.

Where I think garbage time really kicks in is when it's so far gone teams start considering the *season,* rather than the *game* and start pulling guys. That you can't account for.

Points: 0

#8 by ImNewAroundThe… // Nov 22, 2022 - 11:02am

There was 6:18 left when Brissett snapped the ball on 3rd and 11 at their own 24 and only got 6 yards (incompletion the play before). There was a 99.9% Bills win probability at that point. 

Points: 0

#17 by Pat // Nov 22, 2022 - 9:57pm

In late game time compressed situations, a ton of the low probability is due to the low onsides-kick chance, which is a single play. That doesn't mean the other plays are meaningless, it just means the onsides is crazy important. It's part of the reason why switching away from a kickoff is so risky: any alternative to an onsides kick can radically alter late game.

Teams claw back from that situation to where an onsides gives them a realistic chance often. The Browns didn't, because they couldn't force a Bills three and out. That says something about both the Browns and Bills.

Points: 0

#18 by ImNewAroundThe… // Nov 23, 2022 - 12:46am

Pre...facto aka WP. Or post, essentially score (the Bills added 3 onto 25 when the Browns ended at 23). 

It is what it is. It's fine. 

Points: 0

#9 by Paul R // Nov 22, 2022 - 12:20pm

Nitpick: Justin Fields is called "Jones" in the comments under his stat line. 

Would it be useful to add the opponent adjustments to each stat line? Perhaps adding (+25) or something next to the opponent's name in that column? Or would that confuse the issue?

Points: 0

#14 by Vincent Verhei // Nov 22, 2022 - 8:41pm

Nitpick: Justin Fields is called "Jones" in the comments under his stat line. 

Thanks. Fixed. 

Would it be useful to add the opponent adjustments to each stat line? Perhaps adding (+25) or something next to the opponent's name in that column? Or would that confuse the issue?

I'm afraid there's only so much space we have in these tables. A lot of people are reading these on mobile devices and have trouble reading them as it is. I'll note the most extreme adjustments in the player comments. If I don't say anything, you can assume the adjustments are moderate.

Points: 0

#20 by HitchikersPie // Nov 23, 2022 - 6:28am

I know the biggest single season adjustment from DYAR was Brady's '09 with 2,021 DYAR and 1,649 YAR for +372 adjustment, but is there a list of the biggest game adjustments out there, I seem to recall Warner's '09 playoff game vs the Packers being a strong one.

Points: 0

#19 by Theo // Nov 23, 2022 - 2:00am

If you run the exact same thing on every down and distance, and the opponent knows your playcalling, then you can throw it to your HB for a td, and again, and again. 

The Steelers have gone stale. And without Ben pulling games out of his ass, this is just exposing how bland and simple their schemes are. 

Points: 0

Save 10%
& Support Vincent
Support Football Outsiders' independent media and Vincent Verhei. Use promo code VINCE to save 10% on any FO+ membership and give half the cost of your membership to tip Vincent.