Tyler Huntley, Justin Fields, and the QB Rushing Quandary

Baltimore Ravens QB Tyler Huntley
Baltimore Ravens QB Tyler Huntley
Photo: USA Today Sports Images

NFL Week 13 - Quarterback runs: they're exciting, effective, and hard to evaluate. Why does Tyler Huntley (10 carries for 41 yards and a touchdown) lead quarterbacks in rushing value this week instead of Justin Fields (five for 72 and a score)? It's a tricky question to answer, but one we need to address as quarterbacks are running more these days than they ever have before.

Going into Monday Night Football, quarterbacks have run the ball 1,398 times in 2022 for 8,197 yards, averaging 3.59 carries and 21.0 yards per team per game, and 5.86 yards per carry. (We have removed kneeldowns from all rushing statistics discussed in this essay.) By volume, those are the highest numbers in our DVOA database going back to 1981, but the efficiency is par for the course. Here's a look at how quarterback rushing numbers have changed over the past four-plus decades:

QB Runs by Decade, 1981-2022
Decade Runs/G Yds/G Avg TD/G DYAR/G
1980s 1.82 10.47 5.75 0.11 1.09
1990s 1.89 11.06 5.85 0.09 1.87
2000s 1.95 11.20 5.74 0.10 1.80
2010s 2.32 13.93 5.99 0.12 2.21
2020s 3.37 18.95 5.62 0.21 1.84
Kneeldowns not included.
Numbers are per team, per game.

For about 30 years, quarterback rushing totals held steady, but they began to creep upwards in the 2010s before exploding in recent years. The tipping point actually came five years ago. In the year 2000, quarterbacks averaged 2.52 carries per team per game, the most in any season from 1981 to 2017. In 2018—Lamar Jackson's rookie year, not coincidentally—that rate climbed to a new high of 2.80 and has not dropped back to pre-Lamar levels since, with marks of 2.68, 3.35, and 3.24 in the ensuing years before this season surpassed them all.

(A brief aside to discuss the absurd numbers of that 2000 season: quarterbacks that year averaged 6.43 yards per carry, the second-highest rate in our 42-year database. And there were a lot of them, 17 with at least 30 carries, most until 2020. Led by Donovan McNabb, Rich Gannon, Daunte Culpepper, Kordell Stewart, and Jeff Garcia, quarterbacks that year combined for 1,866 rushing DYAR; they have never totaled even 1,500 in any other season.)

For all that volume, however, efficiency on quarterback runs is going down. Through (almost) three years in the 2020s, the average gain on quarterback runs is at an all-time low. Touchdown rates have gone up, but that just means quarterbacks are running near the goal line more often. Game-for-game, rushing DYAR numbers are right in line with what we saw in the 1990s and 2000s, and down sharply from where they were in the 2010s. And carry-for-carry, DYAR numbers have never been lower—quarterbacks are averaging 0.55 DYAR per run so far in the 2020s, lowest for any decade on record. Defenses have caught on, and quarterbacks no longer have the element of surprise. Thus, quarterback runs are not as effective as they used to be.

But they are still more effective than most other football plays—by a lot. The average team offense DVOA on quarterback runs this year is 27.0%. Keep in mind that the Kansas City Chiefs lead the league with an offensive DVOA of 25.0%, meaning the average quarterback run gets better results than the best offense in the NFL. That's what happens when you gain 5.96 yards per play with a 60% success rate.

That's oversimplifying things, however, because not all quarterback runs are created equal. They can be broken down into three categories:

  • the cringeworthy humiliation of aborted snaps and handoffs (-328.5% team offense DVOA, -1.89 yards per play, 6.2% success rate in 2022);
     
  • the brutal reliability of designed quarterback runs such as sneaks, draws, or option keepers (25.3% DVOA, 4.97 yards per play, 65.6% success rate);
     
  • and the Krakatoa-level explosive threat of quarterback scrambles, which count as pass plays in our team offense stats but rushing plays in our in our analysis of individual quarterbacks (73.3% DVOA, 7.60 yards per play, 59.8% success rate).

So when a quarterback runs, his rushing DYAR and DVOA are compared to all of these—to Baker Mayfield dropping a snap (four times in his seven games), to Jalen Hurts pushing forward on a sneak (Hurts leads all quarterbacks with 80 designed runs), and to Justin Fields tucking and running away from edge rushers (Fields has 54 scrambles, a dozen more than anyone else). It's an odd mix that leads to some results that are non-intuitive, to say the least. Geno Smith gets negative DYAR for a 13-yard run on second-and-25 against the Lions, for example, while Patrick Mahomes gets negative DYAR for a 6-yard run on first-and-10 against the Chargers. That's partly due to opponent adjustments—Detroit and L.A., after all, stink—but mostly because the average baselines on quarterback runs are so high.

Which brings us back to Tyler Huntley and Justin Fields. Huntley's longest run against the Denver Broncos gained only 14 yards, but each of his 10 carries gained at least 2 yards and five resulted in first downs, including a 2-yard touchdown run and three other conversions with 2 yards or less to go. Four of his runs had negative DYAR, the worst of which was a 2-yard gain on second-and-5. Fields, meanwhile, had a 55-yard touchdown run in the first quarter, but only one of his other four carries (a 1-yard gain on third-and-1) picked up a first down. Two of his runs had negative DYAR, including a 1-yard gain on second-and-11. To put it in baseball terms, Huntley's singles and walks add up to more value than Fields' home run.

Regardless, both Huntley and Fields should probably get more credit for their rushing contributions than we're actually giving them right now. This is especially true for high-volume, medium-efficiency runners such as Hurts. Hurts is second among quarterbacks with 117 carries, but he actually has a negative DVOA on those plays. He ranks ninth in DYAR, below Jacoby Brissett, Patrick Mahomes, and Joe Burrow, three players with only 124 carries between them. Since quarterback runs on the whole are so valuable, however, we should be doing more to reward Hurts for getting so many carries in the first place. This wasn't a huge issue for most of Football Outsiders history—if your team didn't have Michael Vick or Tim Tebow or Cam Newton, you probably weren't concerned about quarterback rushing DYAR—but as quarterbacks run more and more, it's a problem we will have to address one of these offseasons to keep up with a changing sport.

In the meantime, FO editor-in-chief Aaron Schatz has found a duct-tape-and-chicken-wire solution. Analyzing EPA data on nflfastR, he found that 1 rushing DYAR was worth about 2.1 passing DYAR. So when ranking quarterbacks, we should be multiplying rushing DYAR by 2.1, then adding that total to passing DYAR. This would not have a huge impact on the rankings this week—Daniel Jones and Ryan Tannehill would each move up two slots, while Dak Prescott, Kenny Pickett, Mac Jones would fall two or three, with everyone else within one spot of where they are now. In full-season numbers, electric runners such as Fields, Jackson, and Josh Allen get a boost, while the clumsy and the old (Mayfield, Matt Ryan, Tom Brady) get dinged.

Regardless, if you have been reading Quick Reads in recent years and concluded we were underrating running quarterbacks, you're probably right.

Quarterbacks
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
1.
Jalen Hurts PHI
29/39
380
3
0
2
207
187
20
TEN
If you saw A.J. Brown's highlights, it will not surprise you to learn that Hurts led all quarterbacks in DYAR on deep passes (5-of-9 for 148 yards and three touchdowns, plus a 24-yard DPI) and on throws to the outside (21-of-30 for 265 yards and two scores, plus that 24-yard DPI).
2.
Joe Burrow CIN
25/31
286
2
0
1
185
172
13
KC
Burrow was the best passer on throws to running backs, going 7-of-8 for 57 yards and a touchdown.
3.
Jared Goff DET
31/41
340
2
0
2
175
175
0
JAX
Goff loses 48 DYAR due to opponent adjustments. He still led the league in DYAR on passes to receivers at or behind the line of scrimmage (8-of-10 for 86 yards) and in the first quarter (8-of-9 for 101 yards and a touchdown). His 68% success rate was also best in the league.
4.
Josh Allen BUF
22/33
223
2
0
2
168
160
8
NE
Allen gains 45 DYAR due to opponent adjustments. He was the week's best passer on third/fourth downs (9-of-11 for 87 yards and seven conversions, including a touchdown, plus a 22-yard DPI) and on throws to his right (9-of-12 for 114 yards and a touchdown, plus that 22-yard DPI).
5.
Andy Dalton NO
20/28
229
1
0
2
107
111
-4
TB
6.
Geno Smith SEA
28/39
367
3
1
4
106
107
-1
LAR
Smith was the week's worst passer on throws to receivers at or behind the line of scrimmage (5-of-9 for 30 yards with an interception), but the best on throws to wide receivers (19-of-23 for 273 yards and two touchdowns).
7.
Patrick Mahomes KC
16/27
223
1
0
2
71
60
11
CIN
Mahomes' average completion gained a league-high 9.4 yards after the catch.
8.
Tyler Huntley BAL
27/31
187
0
1
2
59
38
22
DEN
Huntley came into the game with Baltimore down 3-0 in the first quarter. He gains 53 DYAR due to opponent adjustments. He led the Ravens to a come-from-behind win, but he was still the week's worst passer out of a no-huddle, going 1-of-2 for 2 yards with an interception.
9.
Russell Wilson DEN
17/22
189
0
0
2
57
49
8
BAL
If you're wondering how a quarterback can rank this high when his team only scored nine points, I'll point out that Wilson threw for seven first downs, and all of them came on his side of the field. In Ravens territory, he went 4-of-7 for only 19 yards. He also only had one third-down conversion all day, going 5-of-7 for 44 yards with a pair of sacks. But hey, on first and second down on his own side of the field, he was real good.
10.
Dak Prescott DAL
20/30
170
3
1
1
57
57
0
IND
When the third quarter ended, the Cowboys were nursing a narrow 21-19 lead. And then Prescott was the week's top passer in the fourth quarter and overtime, completing all four of his passes for 42 yards and a touchdown,and adding a 44-yard DPI. He was also tops in the red zone, going 8-of-9 for 66 yards and three touchdowns.
11.
Daniel Jones NYG
25/29
200
1
0
4
55
34
21
WAS
Not counting a spike to stop the clock at the end of the first half, Jones completed each of his first 17 passes for 136 yards and a touchdown. Now he did mix a pair of sacks in there, and only four of those completions picked up first downs, which brings up our next point: Jones led the league with 12 failed completions. Still he was the week's best passer in the second quarter, when he completed all seven of his passes for 86 yards.
12.
Tom Brady TB
36/54
281
2
1
1
47
47
0
NO
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
13.
Justin Fields CHI
20/25
254
0
2
0
45
28
17
GB
Fields averaged 10.2 yards per dropback, best in the league—and that's not even including his rushing stats. Unfortunately, like Russell Wilson, he failed to throw for a single first down in opponent territory. After crossing midfield, he went 5-of-8 for 26 yards with two interceptions.
14.
Marcus Mariota ATL
13/24
167
1
1
1
34
25
9
PIT
Mariota's average dropback came with 10.7 yards to go for a first down, most in the league. Perhaps that's why his 15.8-yard average depth of target was also most in the league. But throwing deep often is not the same as throwing deep well—Mariota threw eight deep passes, completing one to Drake London for 37 yards and one to a Steelers defender for an interception.
15.
Kirk Cousins MIN
21/35
173
1
0
2
34
27
7
NYJ
Cousins' average completion gained a league-low 2.1 yards after the catch.
16.
Kenny Pickett PIT
16/28
197
1
0
0
32
45
-13
ATL
Pickett loses 41 DYAR due to opponent adjustments. He was still the week's best passer on throws to tight ends, going 5-of-7 for 101 yards and a touchdown.
17.
Aaron Rodgers GB
18/31
182
1
0
0
26
26
0
CHI
Rodgers loses 50 DYAR due to opponent adjustments. He was at his best on throws down the middle, going 8-of-10 for 93 yards.
18.
Ryan Tannehill TEN
14/22
141
1
0
6
22
5
17
PHI
Tannehill gains 48 DYAR due to opponent adjustments. His dropbacks averaged 3.8 yards apiece and had a 25% success rate, both the worst of any qualifier this week. He had a bad day on third downs, going 3-of-5 for 40 yards with one touchdown, one other conversion, and five sacks.
19.
Derek Carr LV
16/30
250
2
1
0
19
19
0
LAC
So remember last week, when Carr was best in DYAR on throws to the outside but last on throws down the middle? Well apparently he over-corrected, because this week he was the best passer on throws down the middle (completing all five of his throws for 112 yards and a touchdown), but the worst on throws to the outside (11-of-25 for 138 yards with one touchdown and one pick-six).
20.
Mac Jones NE
23/36
195
1
0
1
15
28
-13
BUF
Jones only threw for four first downs against Buffalo, and three of them came with the Patriots trailing by 17 points in the fourth quarter. In the first three quarters, he had one 48-yard touchdown, but his other 11 passes resulted in six completions for only 38 yards.
21.
Brock Purdy SF
25/37
210
2
1
3
10
7
3
MIA
Purdy came into the game with San Francisco trailing 7-3 in the first quarter. His average pass attempt traveled only 5.8 yards beyond the line of scrimmage, lowest in the league, and for good reason: he was the week's worst passer on deep balls, going 0-for-3 with an interception.
22.
Mike White NYJ
31/57
369
0
2
1
3
-4
6
MIN
The Jets lost this game by only five points, which makes it especially painful that White had the week's worst DYAR in the red zone, throwing 17 passes (!!!) inside the Minnesota 20 but completing only six of them for 24 yards, with no touchdowns, one first down, and one interception.
Rk
Player
Team
CP/AT
Yds
TD
INT
Sacks
Total
DYAR
Pass
DYAR
Rush
DYAR
Opp
23.
Tua Tagovailoa MIA
18/33
295
2
2
3
-26
-26
0
SF
Tagovailoa gains 51 DYAR due to opponent adjustments. He still finished with the NFL's worst passing DYAR on third/fourth downs (1-of-7 for 9 yards with one conversion and two sacks) and in the third quarter (3-of-5 for 35 yards with two interceptions).
24.
Trevor Lawrence JAX
17/31
179
1
0
2
-32
-40
9
DET
Lawrence was best on throws to his left, going 7-of-9 for 96 yards and a touchdown.
25.
Deshaun Watson CLE
12/22
131
0
1
1
-44
-41
-3
HOU
Watson's average dropback came with a league-low 7.3 yards to go for a first down. He had the week's worst DYAR on throws down the middle (2-of-6 for 20 yards with an interception) and to wide receivers (8-of-14 for 96 yards with an interception).
26.
Taylor Heinicke WAS
27/41
275
2
0
5
-44
-47
2
NYG
After a good start, Heinicke had the week's worst DYAR after halftime, when he went 16-of-25 for 165 yards and four sacks.
27.
Matt Ryan IND
21/37
233
2
3
3
-63
-66
3
DAL
Ryan gains a league-high 72 DYAR due to opponent adjustments. Remember when we said this was a close game after three quarters, and then Dak Prescott became the league's best passer? Well, Ryan was the league's worst passer in the fourth quarter, when he went 9-of-15 for 62 yards with an 18-yard DPI, one sack, one fumble, and two interceptions.
28.
John Wolford LAR
14/26
178
0
2
4
-68
-65
-3
SEA
The Rams and their skeleton crew somehow lost this game by only four points, and they might have even won if Wolford had played better in scoring range. He was the week's worst passer inside the opponents' 40-yard line, going 6-of-10 for 47 yards with two sacks, one fumble, and one interception.
29.
Justin Herbert LAC
28/47
335
1
0
5
-76
-47
-29
LV
Herbert loses a league-high 90 DYAR (not a typo) due to opponent adjustments. The Raiders only have 21 sacks all season; five of them (24%) came in this game. Herbert is also the first quarterback this year to complete less than 60% of his passes against Las Vegas.
30.
Kyle Allen HOU
20/39
201
1
2
0
-133
-93
-40
CLE
When a team is playing a backup or untested quarterback, they will often try to make his life easier by drawing up short, easy passes to running backs and tight ends. That didn't work for Houston, because Allen was the week's worst passer on throws both to running backs (5-of-7 for 35 yards with a pick-six) and to tight ends (7-of-14 for 67 yards with an interception).
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
12
91
2
2/3
15
0
51
42
9
IND
The Colts stuffed Pollard only once while allowing him to run for four first downs, including a 30-yard touchdown and gains of 17 and 22 yards. His best catch was a 10-yard gain on second-and-7.
2.
AJ Dillon GB
18
93
1
3/3
26
0
41
26
15
CHI
Dillon rushed for five first downs with gains of 16 and 21 yards while being stuffed just once. Two of his catches also gained first downs, and the other was a 5-yard gain on first-and-10.
3.
D'Andre Swift DET
14
62
1
4/6
49
0
39
22
17
JAX
Swift only ran for two first downs and his longest run gained just 10 yards, but he was stuffed only once and half of his runs gained 4 yards or more. All four of his receptions produced first downs, including gains of 14 and 24 yards.
4.
Isaih Pacheco KC
14
66
1
2/2
16
0
36
29
7
CIN
Pacheco was stuffed just once while running for five first downs, the longest a gain of 16. His longest catch also gained 16 yards.
5.
Zonovan Knight NYJ
15
90
0
5/5
28
0
27
14
13
MIN
Who? An undrafted free agent out of NC State with one of the league's coolest names, Knight played his second NFL game and made his first start against the Vikings on Sunday. He had a boom-or-bust day on the ground against Minnesota, being stuffed four times with only two first downs, but one of those first downs was a 48-yard gain in the fourth quarter. He added one more first down as a receiver on an 8-yard gain on second-and-7.
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.
Tony Pollard DAL
12
91
2
2/3
15
0
51
42
9
IND
2.
Isaih Pacheco KC
14
66
1
2/2
16
0
36
29
7
CIN
3.
AJ Dillon GB
18
93
1
3/3
26
0
41
26
15
CHI
4.
Devin Singletary BUF
13
51
1
0/2
0
0
11
24
-13
NE
The Patriots stuffed Singletary twice while allowing him to run for a trio of first downs, the longest a gain of 11. That, with a boost of 10 DYAR in opponent adjustments, is enough to get him on this table in a quiet week for running backs. Josh Jacobs and Samaje Perine were the only backs to run for over 100 yards this week, and they did it against lousy Chargers and Chiefs run defenses, which knocks them out of the top five.
5.
Dalvin Cook MIN
20
86
1
2/3
-3
0
3
23
-20
NYJ
Cook gets a boost of 12 DYAR due to opponent adjustments. The Jets stuffed him five times while allowing him to run for four first downs, including gains of 11 and 21 yards.
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.
Derrick Henry TEN
11
30
0
2/3
8
0
-27
-22
-5
PHI
Well, the good news is that each of Henry's 11 carries gained at least 1 yard. The bad news is that none of them gained more than 6 yards or picked up a first down, and only two counted as successful plays. None of his catches picked up first downs either.
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.
Derrick Henry TEN
11
30
0
2/3
8
0
-27
-22
-5
PHI
Five Best Wide Receivers and Tight Ends by DYAR
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
Opp
1.
Amon-Ra St. Brown DET
11
12
114
10.4
2
69
JAX
St. Brown's totals include 1 rushing DYAR for his one carry, a 1-yard gain. Though his longest catch gained only 15 yards, nine of them picked up first downs, and he added a 10th first down on a 5-yard DPI.
2.
Davante Adams LV
8
12
177
22.1
2
69
LAC
Seven of Adams' eight receptions picked up a first down, including 31- and 45-yard touchdowns, as well as a gain of 36.
3.
DK Metcalf SEA
8
8
127
15.9
1
66
LAR
Each of Metcalf's eight receptions counted as a successful play and five gained first downs, including gains of 25 and 40 yards, not to mention the game-winning 8-yard touchdown.
4.
A.J. Brown PHI
8
10
119
14.9
2
60
TEN
Each of Brown's eight receptions counted as a successful play and five gained first downs, including touchdowns of 29 and 40 yards, and he added a 24-yard DPI.
5.
Christian Watson GB
3
6
48
16.0
1
59
CHI
Watson's totals include 33 rushing DYAR for his one carry, a 46-yard touchdown. His first three targets all fell incomplete, but his next four all resulted in first downs, including a 14-yard touchdown on fourth-and-4 and a 38-yard DPI.
Worst Wide Receiver or Tight End by DYAR
Rk
Player
Team
Rec
Att
Yds
Avg
TD
Total
DYAR
Opp
1.
Zay Jones JAX
2
7
16
8.0
0
-40
DET
The first five passes thrown Jones' way were all incomplete. His two receptions—a 10-yard gain on third-and-11 and a 6-yard gain on first-and-10—both came with Jacksonville down by 24 points in the second half.

Comments

59 comments, Last at 07 Dec 2022, 11:46am

#1 by HitchikersPie // Dec 06, 2022 - 4:04am

I think plenty of us could have predicted that Mahomes would be tops by DYAR through week 13, but I doubt many had the next two in order:

 

1. P. Mahomes 1,281

2. T. Tagovailoa 1,144

3. J. Goff 881

Points: 0

#3 by Joey-Harringto… // Dec 06, 2022 - 5:28am

The first name needs no explanation of course.  I haven’t watched enough Dolphins games to comment on the second name (but Hill and Waddle are really good).  As for the third name, the answer is that Amon-Ra St. Brown is awesome and we should talk about him more.  I heard a stat that Jared Goff is 4th in EPA/drop back when St. Brown has been healthy, and something like 18th when he wasn’t (in the middle of their 5 game losing streak).

Points: 2

#4 by coboney // Dec 06, 2022 - 6:05am

The other part to the answer is that Jared Goff is a competent quarterback when not pressured and the Lions offensive line is very good which means you get a lot more competent Goff, and a lot less of Bad Goff. Essentially, with Goff, the Lions are doing what the Colts hoped to do with Matt Ryan - provide a solid line and weapons to get the best out of a middling QB.

Points: 7

#10 by Joey-Harringto… // Dec 06, 2022 - 8:31am

That's the thing.  Keeping a good line together long term is hard, and injuries are inevitable (as the Colts and Rams found out this year).  Decker and Ragnow are signed long-term, and Sewell is still on his rookie deal, so that's a problem to worry about for another day.  Even then, pressure is inevitable.  Goff still does some galaxy-brained stuff under pressure (like 360 spin near interception against the Bills).  And this is Goff playing at his absolute ceiling.

If the Texans somehow claw out two more wins (unlikely, I know), and that Rams pick goes to #1, I will be jumping up and down for the Lions to take Bryce Young. 

Points: 1

#14 by Aaron Brooks G… // Dec 06, 2022 - 9:36am

It's not like the Lions line has been super healthy. Vaitai and Kraemer have missed the entire season, and Ragnow and Brown have been gimping around a little.

They've basically been down 1.3 starters all year, plus their swing guard. They've been using their backup center as their plug-in lineman.

2018 Goff was a little better, but he was also on a much better team.

Points: 2

#25 by theslothook // Dec 06, 2022 - 11:35am

This was summarized so beautifully, especially this part.

And this is Goff playing at his absolute ceiling

 

That I think is the salient point. Goff may look much worse at times, but he's overall an ok player.  But what he's doing now doesn't feel sustainable.

Points: 2

#2 by HitchikersPie // Dec 06, 2022 - 4:13am

In the meantime, FO editor-in-chief Aaron Schatz has found a duct-tape-and-chicken-wire solution. Analyzing EPA data on nflfastR, he found that 1 rushing DYAR was worth about 2.1 passing DYAR. So when ranking quarterbacks, we should be multiplying rushing DYAR by 2.1, then adding that total to passing DYAR.

Is this true for this year, or generalised from 1999-2022, would be a fun one to boost up the career totals of some QBs (Steve Young and Joe Montana are mostly tied but this could give a sizeable bump to Young):

 

Montana Pass DYAR: 12,080
Montana Rush DYAR: 361
Montana P-R DYAR: 12,441

Montana Rush Adjusted DYAR: 758.1
Montana P-RA DYAR: 12,838.1

Young Pass DYAR: 10,302
Young Rush DYAR: 1,305
Young P-R DYAR: 11,607

Young Rush Adjusted DYAR: 2,740.5
Young P-RA DYAR: 13,043

Points: 5

#5 by BlueStarDude // Dec 06, 2022 - 7:59am

I think the adjustment would need to vary by season, but would also be interested to hear what Vince/Aaron has to say.

Points: 0

#42 by Vincent Verhei // Dec 06, 2022 - 3:41pm

The 2.1 number was calculated with data from the last couple of seasons. I would guess it would be similar for other years, but hopefully we'll have time to re-calibrate QB rushing data this offseason and it won't be worth worrying about. 

Points: 1

#6 by Aaron Brooks G… // Dec 06, 2022 - 8:04am

Defenses have caught on, and quarterbacks no longer have the element of surprise. Thus, quarterback runs are not as effective as they used to be.

I’m not sure this is actually correct. I think what you are seeing is a bifurcation in the underlying data. QB scrambles are still present and still effective as runs into a vacated pass defense. But you are also seeing true rushes by QBs.

These are planned runs, as opposed to broken pass plays. So the defense is expecting them, but they are still efficient versus a 10-on-11 rush, because the QB is helping.

But DVOA doesn’t know which version it is seeing just from the box score. You need video/scouting to separate them.
 

Points: 1

#9 by BlueStarDude // Dec 06, 2022 - 8:26am

I think you're clearly right. The increase in called QB rushes lowers the apparent effectiveness of all QB rushes which previously were largely only scrambles and sneaks.

FO must have a way to differentiate, though, because in the team data the scrambles count as passing DVOA/DYAR, and even this article breaks down the DVOA for the three subcategories. 

Points: 0

#11 by Pat // Dec 06, 2022 - 8:58am

The play by play differentiates, although it's still a bit of a guess.

There's also the RPO+QB run problem, which is separate as well. Calling something a QB run when it only becomes a QB run if the defense does something is weird, because you don't really know what to compare it against. If the QB gains 2 yards when the RB' would've gained zero and the pass would be incomplete, that's a positive play by the QB.

Points: 1

#15 by BlueStarDude // Dec 06, 2022 - 9:58am

True. But not every positive play can be accounted for—when a QB throws the ball away that's a positive (usually, end-of-half clock issues aside) compared to taking a sack or forcing a bad throw. At a certain point you have to trust in a large enough sample size and applying context to the stat. The goal here has to be getting the QB DYAR/DVOA better, not perfect.* It's also another example of why DVOA will always be a better team stat than player stat.

IIRC, someone here (maybe you?) has pushed for a third category for RPO. If it continues to be more common that would seem to be the way to go.

*If I saw it correctly, the Saints had a called pass play last night where Dalton scrambled a bit and then lateraled to the RB turning it into a rushing play: some things are OK not being correctly accounted for, as long as they don't happen too often.

 

Points: 0

#16 by Aaron Brooks G… // Dec 06, 2022 - 10:03am

IIRC, someone here (maybe you?) has pushed for a third category for RPO. If it continues to be more common that would seem to be the way to go.

Pretty much any option play shares features with an RPO -- there's usually a pass option baked somewhere in there. You can argue draws and play-actions share the common feature of presenting the wrong cue to the defense.

Points: 0

#36 by Pat // Dec 06, 2022 - 2:48pm

I mean, almost any play has a run check built into it, which is why I've argued so hard against even bothering with a run/pass differentiation (now, "short target/deep target" that's totally different). The fact that the analytics community has for so long been like "why are you running so much" just drives me nuts - a lot of times offenses run because they already know their original pass play was going to fail. If you've got 5 seconds left on the play clock, checking to a run for 2 yards is way better than dirting the ball.

The neat thing with RPOs is that you can actually see it. So in those cases you know that it's the defense that's dictating run vs pass rather than the offense.

Points: 0

#54 by Will Allen // Dec 06, 2022 - 9:26pm

This cannot be emphasized ennough. To slightly modify what I've said for decades, playcalling is a function of personnel and alignment, and to opine on playcalling, without fleshing out the context provided by those variables, is just a waste of time.

Points: 0

#18 by colonialbob // Dec 06, 2022 - 10:21am

Defenses have caught on, and quarterbacks no longer have the element of surprise. Thus, quarterback runs are not as effective as they used to be.

 These are planned runs, as opposed to broken pass plays. So the defense is expecting them, but they are still efficient versus a 10-on-11 rush, because the QB is helping.

Both of these can be true, though - designed QB rushes can be more effective than RB rushes (because of the numbers advantage you mention), while still declining in effectiveness (at least by efficiency) as they are used more often, since defenses begin to plan around them and develop counters to them. 

Points: 0

#22 by Aaron Brooks G… // Dec 06, 2022 - 11:27am

Even at the NCAA level, option rushing is more efficient than standard rushing (because numbers), even though defenses expect it more. 

Scrambles are absurdly effective because they are such a biased sample -- you don't scramble if the defense doesn't give you a beneficial look. A planned rush will never be as efficient as a fortunate scramble.

You see it almost never at the NFL level, but this is in part why option passing is also absurdly efficient. There's almost always a rush option baked in, so you can scrub it if the defense doesn't take the bait. This is how Demaryius Thomas had 1150 receiving yards and 8 TDs on 46 receptions in an offense that threw the ball 160 times total. (They ran about 700 times)

As it turns out, the fortunate pass is also more efficient than the planned pass.

Points: 0

#53 by Pat // Dec 06, 2022 - 8:02pm

"As it turns out, the fortunate pass is also more efficient than the planned pass."

Also known as the 'defense done screwed up' pass.

Points: 0

#32 by JimZipCode // Dec 06, 2022 - 1:54pm

Defenses have caught on, and quarterbacks no longer have the element of surprise. Thus, quarterback runs are not as effective as they used to be.
----------------------
I’m not sure this is actually correct.

I'm not sure either, but it matches the eye test from watching Ravens game.  Defenses are extraordinarily alert to Lamar running; areas of the field that look wide open, collapse very fast.  Seems like gaps that would've been ~20 yard gains three years ago, are now like ~6-8 yard gains.  Everything comes much harder.

That Ravens have not adjusted well, in terms of personnel to threaten outside and force defenses to respect the pass, so there's that. 

Points: 0

#34 by theslothook // Dec 06, 2022 - 2:10pm

Anecdotally, this seemed to happen to Vince Young and Michael Vick as well. Over time, especially the division rivals, got better at defending their run style. 

Points: 0

#52 by whocares4 // Dec 06, 2022 - 7:17pm

Is that true? Vick's best rushing year was his final year with the Falcons (2006) according to both DYAR and conventional stats (though 2004 is close.) When he was a rushing QB in his 6th year in the league, defenses don't seem to have figured him out at all.

When he went to the Eagles, Andy Reid trained him to be a real QB and he no longer relied on the run. His passing DVOA/DYAR went from being consistently horrible (32nd/32nd in his last year in Atlanta) to consistently good (7th/14th in his first year in Philadelphia) - he was a different player in the two ears of his career so I'm not sure this theory holds up. If his rushing production declined, it's purely because he became an actual passer and not a "single read and take off" guy.

Points: 0

#7 by Aaron Brooks G… // Dec 06, 2022 - 8:10am

To put it in baseball terms, Huntley's singles and walks add up to more value than Fields' home run.

Baseball analysts would be horrified by that statement.

Points: 1

#17 by speaker42 // Dec 06, 2022 - 10:03am

Not at all.  To use the most crude approach to advanced metrics, the two big things a hitter does for you are get on base (OBP, reflected in hits and walks) and hit for power (SLG, reflected primarily in home runs and doubles).  Of course one home run is more valuable than one single or one walk.  But a 2-for-3 performance with 2 singles and a walk can outweigh a 1-for-4 performance with a home run, because the home run hitter has exhausted more of the game's most precious commodity:  outs.

TTO players continue to get playing time because two of the outcomes (home runs and walks) are positive.  If they didn't have those walks, they'd be pinch-hitting specialists, if they still had jobs at all.

Here, getting on base is analogous to acquiring new sets of downs, while hitting a home run is analogous to scoring a touchdown.  A long touchdown run is obviously more valuable than a 2-yard run in the middle of the field, even if that 2-yard run picks up a first down.  But a bunch of first-down runs can generate more value overall.

[And yes, I know you know all this and just made a flippant comment for fun.  But I just...can't...help myself sometimes.]

Points: 5

#23 by Aaron Brooks G… // Dec 06, 2022 - 11:32am

And annoyingly, the Indians continue to be good and the Royals have that WS win (pennants fly forever).

I'm not sure the death of the single has been good for baseball. It does amuse me that DVOA hates the football equivalent of the home run, though, and basically treats it as a double. But it *loves* RBIs coming from a man on 3rd and two outs.

Points: 3

#27 by speaker42 // Dec 06, 2022 - 12:18pm

There is no longer a MLB team with the nickname "Indians."  Baby steps.

The 1980s St. Louis Cardinals model is absolutely still a viable offensive strategy.  But you can't stop stealing bases when you have a big lead, and you certainly shouldn't be giving up outs to move players from first base to second early in the game.  The path to success is narrower, but when everyone is dropping wads of cash on big, slow, powerful defensive liabilities, the market inefficiency moves back to the previously heralded .300 hitter.  It's just that, if all your hitter does is hit .300 (no power, no walks), that's not really a sustainable approach - and that's not a market inefficiency, it's a trap.

Points: 1

#38 by JonesJon // Dec 06, 2022 - 3:18pm

I think the problem with the comparison (and the football formula as a whole) is that it doesn't scale the same way. If player A goes 1-4 with a HR every game and player B goes 2-3 with 2 singles and a walk every game player B will wind up with more hits, walks and a significantly higher average, OBP, and OPS. Without knowing any advanced stat formula we can see that a player who hits 667/750/667 is better than one hitting 250/250/1.000.

Now compare Fields and Huntley if you take those numbers for a full season. Fields runs for over 500 more yards, over 10 yards more per carry, the same number of TDs, and a lot less first downs. In a way the argument is almost the opposite, we're being asked to accept that first downs are so valuable that they can wipe away that every other aspect of Fields' line ranges from significantly better to equal.  

Points: 0

#28 by takeleavebelieve // Dec 06, 2022 - 12:32pm

I appreciate your confidence, but this is factually wrong. 2 for 3 and a walk is more valuable than 1 for 4 with a HR using linear weights. 

https://library.fangraphs.com/principles/linear-weights/

Points: 0

#29 by Aaron Brooks G… // Dec 06, 2022 - 1:32pm

You can find any weighting that supports pretty much any argument.

If I scale the Hundley and Fields examples based on the above-discussed examples -- Hundley had 6 successes in 10 chances, but all were singles or walks, whereas Fields had a home run and a single in five at bats. So Hundley was basically 2/4 with a walk, whereas Fields was 2/5 with a home run.

Fields' OPS is 1.4 (2.35 expected runs) vs Hundley's 1.1 OPS (1.95 expected runs). (It's a useful rule of thumb that two home runs (and three outs) is worth about five singles.) Fields had a better day in baseball terms.

Points: 0

#40 by takeleavebelieve // Dec 06, 2022 - 3:29pm

Linear Weights change with the run environment; it’s true that a HR in 1931 might not be weighted the same as an HR in 1968 or 2022, but that’s irrelevant to this discussion. OPS is also completely irrelevant to LWs - the entire point of LWs that equal OPSes aren’t equally valuable.

Maybe instead of inventing arbitrary values, you could go find the actual values that support your argument? What even is a “single” in football terms, and why isn’t it a double? 

After all, it’s your argument. It's your burden to prove it.

\spoilers: you can't prove it

Points: 1

#46 by Aaron Brooks G… // Dec 06, 2022 - 4:35pm

The single is taken from the FO summary. I treated it as equivalent to a successful rush.

Which brings us back to Tyler Huntley and Justin Fields. Huntley's longest run against the Denver Broncos gained only 14 yards, but each of his 10 carries gained at least 2 yards and five resulted in first downs, including a 2-yard touchdown run and three other conversions with 2 yards or less to go. Four of his runs had negative DYAR, the worst of which was a 2-yard gain on second-and-5. Fields, meanwhile, had a 55-yard touchdown run in the first quarter, but only one of his other four carries (a 1-yard gain on third-and-1) picked up a first down. Two of his runs had negative DYAR, including a 1-yard gain on second-and-11. To put it in baseball terms, Huntley's singles and walks add up to more value than Fields' home run.

Huntley had 10 attempts, with six successes (four had negative DYAR). These were treated as "singles." Fair, I think, given his TD was 2 yards. That's like walking in a run.

Fields was two successes in five attempts, but one was a HR.

So normalizing Huntley to 5 plate appearances to match Fields, I have him credit for two singles and a walk (3 successes) in five plate appearances: 2/4 with a BB; both hits are singles.

Fields got 2/5 with a HR. A 55 yard seems appropriate to score as a HR, as the original argument does.

By both OPS and LRS, Fields had the better day. If you would like to argue Huntley was actually hitting doubles, that's an argument against Vince, not against me.

Points: 0

#48 by Vincent Verhei // Dec 06, 2022 - 4:57pm

The point is that Huntley's many small plays were worth more than Fields' one big one.

Points: 1

#50 by Aaron Brooks G… // Dec 06, 2022 - 6:27pm

That's because DVOA treats home runs as though they were doubles. =)

\'course, Aaron grew up w/ Fenway...

Points: 1

#49 by takeleavebelieve // Dec 06, 2022 - 5:50pm

1) You still don’t seem to grasp the concept of what Linear Weights actually measure in baseball. 

2) You are “normalizing” because it conveniently suits your argument, despite the fact that it’s not relevant or even mathematically correct to do so.

Again, I applaud your confidence, but your points might land better if you actually understand what you’re taking about. 

Points: 0

#51 by Aaron Brooks G… // Dec 06, 2022 - 6:27pm

Neither are true, but please continue to put words in my mouth. It saves me time.

Points: 0

#55 by takeleavebelieve // Dec 06, 2022 - 11:16pm

My words are based in facts and actually stand up to scrutiny. You should be thanking me.

Points: 0

#59 by theslothook // Dec 07, 2022 - 11:30am

I don't follow baseball and I hate being the tone police, but I respect both you of as posters. We don't need to be uncivil with one another even if we vehemently disagree. 

Points: 2

#60 by speaker42 // Dec 07, 2022 - 11:46am

Volume matters here too.  6/10 is different than 3/5.  Using up 4 outs is worse than using up 2; but successfully continuing the inning 6 times is better than doing it 3 times.  I'm not prepared to assign relative weights to those things, especially since we're hitting the analogy's breaking point (fresh sets of downs are easier to acquire than additional outs).  But if we assigned Expected Runs to each player, I think it safe to say 6/10 would be better than 3/5.  So artificially deflating Huntley's attempts is changing the parameters of the discussion.

I do think there is a strong argument that 5 (or even 10) plays in a single game is going to attract a lot of noise swamping signal.  Baseball plate appearances are much better candidates for evaluation as independent events.

Points: 0

#8 by Aaron Brooks G… // Dec 06, 2022 - 8:11am

Is the 2.1:1 ratio just true for QBs, or for all rushes?

Points: 0

#43 by Vincent Verhei // Dec 06, 2022 - 3:41pm

Only for QBs. The average QB run is far more efficient than the average RB run.

Points: 1

#12 by Pat // Dec 06, 2022 - 9:07am

Appreciate looking into it, but there's no way the "multiply by 2.1" hack is right. It's a baseline issue, not a scale issue. It might work for combining pass+run crudely.

But it definitely doesn't work for rush comparison. A QB who rushes 10 times for 0 DYAR isn't equivalent to one who doesn't rush at all, even if I multiply it by a bajillion.

It'd be interesting to look at it in a baseline sense, although I confess can't think of a way to do it right now.

Points: 2

#13 by HitchikersPie // Dec 06, 2022 - 9:10am

I think a regression onto some kind of DYAR = constant*EPA+constant could be an easy place to move, you can isolate out just runs from nflindex.com and go from that

Points: 0

#37 by Pat // Dec 06, 2022 - 3:00pm

You'd need to baseline it off of number of carries, though. A QB who runs 10 times for 0 DYAR isn't the same as a QB who hasn't run at all. That's kinda the problem.

Points: 0

#44 by Vincent Verhei // Dec 06, 2022 - 3:43pm

Appreciate looking into it, but there's no way the "multiply by 2.1" hack is right. It's a baseline issue, not a scale issue. It might work for combining pass+run crudely.

Correct. Thus, the "duct tape and chicken wire" comparison. It's a short-term fix to address the problem until we have more time to find a better, permanent solution.

Points: 0

#58 by Pat // Dec 07, 2022 - 9:32am

I feel like you've already got the metric to deal with this: effective yards. I really feel like the problem is that with rushing QBs with high attempts you just don't have a good baseline. But EYds doesn't use a baseline.

Points: 0

#19 by Badfinger // Dec 06, 2022 - 10:41am

Regardless, if you have been reading Quick Reads in recent years and concluded we were underrating running quarterbacks, you're probably right.

 

The quick napkin math on who this affects most this year would put the top QBs as Mahomes/Allen/Burrow/Hurts, Jackson and Brissett move comfortably towards the top 10, and Fields moves into the top 20ish range. These all feel appropriate to me.

However, it makes Daniel Jones a top 15 QB. Are we comfortable with the numbers on this??

Points: 0

#24 by Aaron Brooks G… // Dec 06, 2022 - 11:34am

Someone has to be responsible for those wins.

And it can't be Barkley, because he, *gasp*, runs the ball.

Points: 0

#26 by Will Allen // Dec 06, 2022 - 11:57am

I can't believe those dumb football coaches still use formations with running backs.... .

Points: 0

#30 by Aaron Brooks G… // Dec 06, 2022 - 1:33pm

I can't believe Strunk doesn't just fire GMs who draft them!

\oh wait...

Points: 1

#35 by Badfinger // Dec 06, 2022 - 2:32pm

They already appropriately punished Lamar Jackson in the stats for throwing the ball up on 3rd and Ballgame, does someone else need credit for this??

Points: 0

#39 by JonesJon // Dec 06, 2022 - 3:21pm

Jones is 15th by QBR and 11th in EPA/CPOE composite. So that would put him in line with what some of the other measures are saying

Points: 1

#20 by Raiderfan // Dec 06, 2022 - 11:08am

Herbert should have lost more DYAR; you are overestimating the quality of the Raiders pass defense.

Points: 1

#21 by serutan // Dec 06, 2022 - 11:11am

Oops.  misread.  Nothing to see here.

Points: 0

#31 by BigRichie // Dec 06, 2022 - 1:48pm

However many sacks are failed scrambles. Probably 99% of 0-1 yard sacks.

Saw it happen to Dak Sunday night. Had a clean pocket, but thought he saw an opening so chose to run, then oops! (that was actually a 5-6 yard sack)

Don't know you could even begin to assess this without film study of every single sack. And then it would still often wind up being quite subjective.

Points: 0

#41 by JonesJon // Dec 06, 2022 - 3:40pm

The other side is true too. A good portion of the 1-2 yard scrambles are guys barely running out of sacks

Points: 0

#33 by JimZipCode // Dec 06, 2022 - 1:55pm

Love the thoughtful phrases here:

  • the cringeworthy humiliation
  • the brutal reliability
  • and the Krakatoa-level explosiveness

Fun!

Points: 1

#45 by ChrisS // Dec 06, 2022 - 4:03pm

After listening to many many many broadcasters these QB numbers will all be ignored and we shall rank them by WINS/losses, The One True Statistic. Did you know that Burrows (by himself obviously) is the only QB to ever defeat Mahomes 3 times in a row?

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.