News and commentary from around the web

Football Outsiders content published by ESPN

# Week 2 Quick Reads

by Vince Verhei

NFL analytics Twitter can be a fun place. In August I read this Tweet by Scott Barrett of Pro Football Focus, detailing a conversation with Adam Harstad of Dynasty Theory concerning quarterbacks and how they dealt with pressure. To paraphrase, Harstad described quarterbacking as a "three-legged stool," and how pressure forced quarterbacks to choose between those three legs. When under pressure, quarterbacks must choose which risk to take. Do they hang on to the ball and risk taking a sack? Force a throw into coverage and risk an interception? Or throw the ball away and risk the opportunity to make a play? Good quarterbacks will minimize those risks and make the wise decision in each situation, but sooner or later everyone who throws a pass will be forced to make one choice or another.

Which quarterbacks tend to make which decisions, and is there anything to learn from that the style of football that each quarterback plays? To find out, I took the 55 quarterbacks who threw at least 2,000 passes from 2000 to 2018 (ignoring for now the early portions of 2019, which is still a work in progress). For each of those quarterbacks, I calculated three statistics:

• Sacks per dropback.
• Interceptions per dropback.
• A modified version of yards per attempt, awarding a 20-yard bonus for passing touchdowns, and subtracting interceptions from pass attempts so they are not counted twice.

I then converted each of those stats to a percentage: 100% for the best player in each category, 0% for the worst, with everyone else falling along that line. Next, I charted those three lines in a triangle on an XY scatter plot, with the best players in each category in the middle of the triangle and the worst players in the corners. This gave me three reference points for each player. Finally, for each passer I found the average coordinates of those three points.

The following chart was the result:

Here's how to read this chart: the quarterbacks in green are Gunslingers -- they can move the ball and avoid sacks, but they tend to risk interceptions. As such, it's no surprise to see Brett Favre as the highest name in green. Favre has thrown nearly 60 more interceptions than anyone else in NFL history, and over 90 more than any active player. In this study, Favre has the third-best sack rate and ranks 21st in yards/TDs, but sixth-worst in interception rate. (Remember, these numbers only include seasons since 2000, so a good chunk of Favre's early career is not included.) Other prominent Gunslingers include Kurt Warner and Tony Romo; the most prominent active Gunslinger is Ryan Fitzpatrick.

The players in red are Game Managers. They don't take sacks, and they don't throw interceptions, but they're not particularly good at moving the ball downfield. Sam Bradford (sixth-best interception rate, fourth-worst in yards/TDs) and Derek Carr (top ten in both sack rate and interception rate, 40th in yards/TDs) are the standout players here. (When you look at Carr's per-game numbers, he's basically the same guy as Tom Brady, just without the yards and touchdowns.) Joe Flacco is the next-most prominent active Game Manager.

The final group, in blue, are players I call Bystanders, which is a name I just made up. These are the guys who will stand by in the pocket forever waiting for receivers to get open downfield. As a result, they get big plays, and they avoid interceptions, but sometimes they wait too long and take a sack. Russell Wilson (second in yards/TDs and third in interception rate, but fourth-worst in sack rate) and Aaron Rodgers (third, first, 42nd in the same categories) are the obvious standouts here. This, obviously, is the land of the running quarterback -- Michael Vick is also in this area, while Colin Kaepernick and Robert Griffin would both be down and to the right from Rodgers if they had enough attempts to qualify.

It's important to remember that we're looking at each quarterback's style, not quality. A theoretical quarterback who ranked first in all three categories would land right in the center of this chart, but a quarterback who finished last in all three categories would also land right in the center of this chart. It's also crucial to keep in mind that each data point represents a quarterback's RELATIVE strengths and weaknesses. So yes, Peyton Manning comes out very close to Jake Delhomme and Mark Sanchez, because the biggest weakness for each passer was interceptions. But Manning was still vastly superior to Delhomme and Sanchez; he had the lowest sack rate of any quarterback in this study and ranked fourth in yards/TDs, but was "only" 22nd in interception rate. Delhomme ranked 19th, 20th, and 48th in those categories, while Sanchez ranked 37th, 44th, and dead last. At the same time, we can say that Drew Brees is similar to, but much better than, Matt Schaub. Same goes for Tom Brady and Drew Bledsoe (of all people).

The Brady-Bledsoe connection brings up another point: quarterbacks from the same team tend to cluster together. Favre is joined at the top of the chart by other Jets quarterbacks, including Fitzpatrick, Sanchez, and Vinny Testaverde. They're all just above a pile of Broncos -- Manning, Brian Griese, Jay Cutler, Jake Plummer. Steve McNair lands in a different bin from his ex-teammate Kerry Collins, but the two Titans are still in the same region. That also applies for ex-Jaguars Mark Brunell and David Garrard. Warner is near Trent Green, the man he replaced with the Rams who also spent most of his career with Dick Vermeil.

Not all the connections are team-related. The Manning brothers are very close together. The Carr brothers are near the same horizontal level -- Derek basically took David's sacks and turned them to dumpoffs instead. Daunte Culpepper, Ben Roethlisberger, and Cam Newton -- a trio of athletic giants -- can be found in a diagonal line on the left side of the chart.

I also want to note some players who landed at the extremes, but opposite the players in the corners. First, look at Joey Harrington over there all by his lonesome. That's what happens when you finish with the fifth-best sack rate, but the seventh-worst interception rate and the very worst rate of yards/TDs. In the middle of the bottom of the chart we find Alex Smith. You may be surprised to see him as a Bystander and not a Game Manager, but Smith had the ninth-worst sack rate to go with his top-five interception rate (he was also 36th in yards/TDs). And finally there's Tony Romo in the upper left. Romo ranked in the 20s in both sack rate and interception rate, but he was first overall in yards/TDs. Really! First! Tony Romo! If sacks and interceptions didn't count he would be the best passer of this century, better than Brady, Manning, Brees, and all the rest. Of course, sacks and interceptions do count -- a lot.

It remains to be seen if this information is actually useful or merely trivial. Do players in different groups age differently? Do they tend to move around the chart in a predictable manner? Do players from different schools or coaching staffs produce similar outcomes? These are questions we may be able to answer with further research.

It's way too early to look at these numbers for 2019 (there are nine qualifying quarterbacks who have yet to throw an interception), but just for fun let's look back at this table for the qualifying quarterbacks of 2018 (along with current MVP candidate Lamar Jackson, who threw 170 passes as a rookie in Baltimore):

Ryan Fitzpatrick's absurd season (he finished first in yards/TDs, but last in interception rate) kind of ruins this whole graph for everyone. He's so far out in left field by himself that he pushes more than half the league into the game manager bucket. It's interesting that Rodgers made such a drastic move to the right compared to his career numbers, moving out of the Bystanders category entirely. This may not be good news -- his sack rate of 7.6% was right in line with his career rate of 7.0%, but he threw a touchdown on only 4.2% of his pass attempts, the lowest rate of his career. It's that lack of production that moved him so far. Both of those rates are up slightly through two games of 2019, but he has yet to throw an interception, after throwing only two in all of 2018. Because some things never change.

 Quarterbacks Rk Player Team CP/AT Yds TD INT Sacks Total DYAR Pass DYAR Rush DYAR Opp 1. Patrick Mahomes KC 30/44 443 4 0 2 199 199 0 OAK In the second quarter alone, Mahomes went 12-of-17 for 278 yards and four touchdowns. Going into Monday Night Football, only 12 other quarterbacks had thrown four touchdowns this season. Mahomes' last five throws of the first half: 42-yard touchdown on third-and-20; 32-yard gain; 43-yard gain; 27-yard touchdown on second-and-17; 39-yard touchdown. Those five plays had a total of 26 yards after the catch. 2. Jimmy Garoppolo SF 17/25 297 3 1 0 165 164 1 CIN Garoppolo had five completions that each gained 36 yards or more. Those five plays gained 47 yards through the air and 143 yards after the catch. Two of them were caught behind the line of scrimmage. 3. Tom Brady NE 20/28 264 2 0 2 155 148 7 MIA Throwing down the middle, Brady went 6-of-7 for 96 yards. Five of those completions produced first downs; the other was an 18-yard gain on first-and-20. 4. Dak Prescott DAL 26/30 269 3 1 1 143 120 23 WAS Prescott ranks this high despite finishing next to last in first-quarter DYAR. In the game's first 15 minutes, he went 4-of-6 for all of 19 yards with zero first downs and one interception. 5. Russell Wilson SEA 29/35 300 3 0 4 129 119 10 PIT Wilson was deadly in scoring range. Inside the Pittsburgh 30-yard line he went 4-of-5 for 67 yards. Three of those completions resulted in touchdowns; the other was a 13-yard gain on third-and-10. 6. Lamar Jackson BAL 24/37 272 2 0 2 117 87 30 ARI Fear the Ravens tight ends! Throwing to that position, Jackson went 10-of-13 for 127 yards and two touchdowns. 7. Matthew Stafford DET 22/30 245 2 2 0 95 96 -1 LAC Stafford did not throw a pass in the red zone, but he had a full day's worth of big plays in the front zone between the 20- and 40-yard lines: 4-of-7 for 76 yards and two touchdowns, but also an interception. 8. Jameis Winston TB 16/25 208 1 0 3 63 60 3 CAR A rough day on third downs: 4-of-9 for 24 yards with as many conversions (two) as sacks. 9. Case Keenum WAS 26/37 221 2 0 1 62 62 0 DAL Keenum ripped up the middle of Dallas' defense, going 12-of-17 for 141 yards and a touchdown. He had 10 first downs on throws down the middle, but only three first downs on 20 throws to either side. 10. Josh Allen BUF 20/30 253 1 0 3 58 48 10 NYG Allen was nearly perfect on throws that went medium-range or deeper. On passes that traveled at least 6 yards downfield, he went 12-of-15 for 207 yards and a touchdown; a 16th throw resulted in a 9-yard DPI. Eleven of those completions picked up first downs; the other was an 11-yard gain on third-and-19. 11. Jared Goff LAR 19/28 283 1 0 3 43 36 7 NO The theme this week is "quarterbacks who ripped up the middle of opponent's defenses." Goff's numbers: 7-of-8 for 125 yards and a touchdown. 12. Philip Rivers LAC 21/36 293 0 1 1 42 36 6 DET Apparently Rivers did not get the memo about excelling down the middle, where he went 3-of-8 for 46 yards. Two of those completions went for first downs, but the other went for no gain on second-and-14. Rk Player Team CP/AT Yds TD INT Sacks Total DYAR Pass DYAR Rush DYAR Opp 13. Aaron Rodgers GB 22/34 209 2 0 2 35 35 0 MIN Rodgers was the anti-Dak. He led the NFL in first-quarter DYAR this week, going 9-of-10 for 134 yards and two touchdowns. He had seven first downs in the first quarter, but only six after that, and he was below replacement level in the last 45 minutes. 14. Kyler Murray ARI 25/39 349 0 0 3 34 39 -6 BAL On third and fourth downs, Murray went 6-of-11 for 74 yards, but only three conversions, with one fumbled snap. 15. Joe Flacco DEN 35/50 292 1 1 2 27 27 0 CHI By DYAR, Flacco's best and worst throws came in the red zone. Inside the Chicago 20, he went 5-of-11 for 37 yards with one touchdown and one interception. 16. Jacoby Brissett IND 17/27 146 3 1 3 12 7 5 TEN A weird day, mostly spent stuck in his own end of the field. He only had five plays in Tennessee's territory, and one of those was a sack-fumble. He also threw an interception from his own 48. But he was literally perfect in the red zone -- all three of his passes there resulted in touchdowns, for 3, 4, and 12 yards. 17. Mason Rudolph PIT 12/19 112 2 1 0 11 7 4 SEA Rudolph came into the game with Pittsburgh up 10-7 at halftime; two passes later he threw the interception that set up Seattle's go-ahead score. In Seahawks territory, he went 5-of-6 for 29 yards with a pair of goal-to-go touchdowns. 18. Luke Falk NYJ 20/25 198 0 0 2 -3 -3 0 CLE 19. Mitchell Trubisky CHI 16/27 120 0 0 0 -6 -8 3 DEN Trubisky did not throw a pass in the red zone. He had only three plays in Denver territory, going 2-of-3 for 15 yards. On third and fourth downs, he went 4-of-8 for 43 yards with only two conversions. 20. Baker Mayfield CLE 19/34 325 1 1 3 -12 -12 0 NYJ 21. Matt Ryan ATL 27/43 320 3 3 1 -14 -13 -1 PHI Ryan struggled frequently against Philadelphia, but not in short yardage. With 5 yards or less to go for a first down, he went 6-of-9 for 95 yards and two touchdowns. Each of his other four completions picked up first downs as well. 22. Ben Roethlisberger PIT 8/15 75 0 0 1 -19 -19 0 SEA Roethlisberger played the entire first half before leaving with an elbow injury. On second and third downs, he went 2-of-9 for 16 yards and a sack. Neither of those completions resulted in first downs, though he did get one on a 16-yard DPI. Rk Player Team CP/AT Yds TD INT Sacks Total DYAR Pass DYAR Rush DYAR Opp 23. Cam Newton CAR 25/51 333 0 0 3 -21 7 -28 TB Newton only completed one of his six red zone passes, and that one was an 8-yard gain on third-and-10. On third and fourth downs, he went 5-of-13 for 33 yards with one sack and only three conversions. 24. Teddy Bridgewater NO 18/30 165 0 0 2 -31 -26 -6 LAR Bridgewater entered the game with New Orleans down by three late in the third quarter. His average dropback came with 11.6 yards to go for a first down, most of any quarterback this week. You'll recall that offenses start with 10 yards to go and then, theoretically, that number gets smaller. 25. Andy Dalton CIN 26/42 311 2 1 4 -39 -27 -11 SF Dalton had 21 dropbacks in the second half, each with a deficit of 21 points or more. Only three of those dropbacks resulted in first downs. All told, he went 11-of-18 for 147 yards (a total that is boosted by a meaningless 66-yard touchdown in the game's final minute) with three sacks after halftime. 26. Marcus Mariota TEN 19/26 154 1 0 4 -42 -44 2 IND Mariota only threw one deep pass against the Colts, an incompletion to Delanie Walker on third-and-5 in the fourth quarter. 27. Eli Manning NYG 26/45 250 1 2 1 -49 -50 1 BUF Third and fourth downs: 4-of-12 for 22 yards with three conversions, one sack, one fumble, and one interception. 28. Carson Wentz PHI 25/43 231 1 2 3 -49 -54 4 ATL Wentz's worst struggles came in two areas of the field. The first was between the 40s, where he went 4-of-12 for 30 yards with a sack and two interceptions. The other was in the red zone, where he went 5-of-10 for 24 yards with one sack. Only one of those completions resulted in a first down, a fourth-down score to Nelson Agholor. 29. Gardner Minshew JAX 23/33 213 1 0 4 -59 -82 23 HOU Minshew fumbled on three of his sacks, all of them on first down. He only threw two passes down the middle of the field: a 2-yard gain on second-and-8 and a 10-yard gain on first-and-20. 30. Deshaun Watson HOU 16/29 159 0 0 4 -78 -86 8 JAX Throwing to his left, Watson went 4-of-8 for just 16 yards. 31. Derek Carr OAK 23/38 198 1 2 3 -107 -99 -8 KC Third-quarter passing: 3-of-6 for 28 yards with two interceptions. 32. Kirk Cousins MIN 14/32 230 1 2 1 -122 -113 -9 GB Cousins' first two completions resulted in first downs. He only had four first downs the rest of the day. He did not convert a third down until the Vikings were down by 11 points in the third quarter. Inside the Green Bay 40, he went 1-of-6 for 5 yards with an interception. 33. Josh Rosen MIA 7/18 97 0 1 3 -126 -126 0 NE 28-of-36 for 444 yards with five touchdowns and no interceptions. No, those aren't Rosen's splits in any category. That's what Alabama quarterback Tua Tagovailoa did against South Carolina. Just sayin'. As for Rosen, all of his dropbacks this week came in the fourth quarter with the Dolphins down by at least 37 points. 34. Ryan Fitzpatrick MIA 11/21 89 0 3 4 -209 -214 4 NE Fitzpatrick's second pass of the game was complete for 15 yards on first-and-10. The next time he picked up a first down, the Dolphins were down by 30 points in the fourth quarter. He had three plays in New England territory: two pick-sixes and a third-down sack. He did not convert any of his seven third-down plays, going 1-of-5 (a 2-yard completion on third-and-14) with two interceptions and two sacks. I swear every word of this write-up is true. Fitzpatrick and Rosen combined for -336 DYAR this week, a total that would set the single-game record for negative DYAR (pending end-of-year opponent adjustments) if it had been done by one player.

 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. Raheem Mostert SF 13 83 0 3/4 68 1 72 27 46 CIN All of Mostert's carries gained at least 1 yard; three of them gained 10 or more. All three of his receptions gained at least 13 yards and a first down, the longest a 39-yard touchdown. 2. Dalvin Cook MIN 20 154 1 3/3 37 0 56 43 13 GB Cook was stuffed only three times while rushing for six first downs, including a 75-yard touchdown. Each of his catches gained at least 11 yards, though only one picked up a first down. 3. Matt Breida SF 12 121 0 1/1 11 0 56 49 7 CIN Breida lost 4 yards on one run, but each of his other carries gained at least 4 yards, and five gained 12 yards or more. 4. Aaron Jones GB 23 116 1 4/6 34 0 35 26 10 MIN Jones had five first downs on the ground and 15 runs that gained 4 yards or more, while being stuffed four times. Two of his receptions resulted in first downs. 5. Jordan Wilkins IND 5 82 0 1/1 4 0 27 28 -1 TEN Each of Wilkins' runs gained at least 3 yards, including 15- and 55-yard runs.

 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. Matt Breida SF 12 121 0 1/1 11 0 56 49 7 CIN 2. Dalvin Cook MIN 20 154 1 3/3 37 0 56 43 13 GB 3. Saquon Barkley NYG 18 107 1 3/7 28 0 24 32 -8 BUF Only one of Barkley's runs was stuffed, while five gained first downs and four gained 10 yards or more, including a 27-yarder. 4. Jordan Wilkins IND 5 82 0 1/1 4 0 27 28 -1 TEN 5. Raheem Mostert SF 13 83 0 3/4 68 1 72 27 46 CIN

 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. Christian McCaffrey CAR 16 37 0 2/6 16 0 -35 -30 -5 TB Six of McCaffrey's carries went for no gain or a loss; only two went for a first down, with a long of 10. His two catches both went for first downs.

 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. Christian McCaffrey CAR 16 37 0 2/6 16 0 -35 -30 -5 TB

 Five Best Wide Receivers and Tight Ends by DYAR Rk Player Team Rec Att Yds Avg TD Total DYAR Opp 1. Demarcus Robinson KC 6 6 172 28.7 2 94 OAK Five of Robinson's catches produced first downs, and four gained 28 yards or more, including touchdowns of 39 and 44 yards. 2. Chris Godwin TB 8 9 121 15.1 1 55 CAR Five of Godwin's catches produced first downs, and four gained 20 yards or more. 3. Davante Adams GB 7 9 106 15.1 0 53 MIN Each of Adams' catches produced a first down, including gains of 21 and 39 yards. In addition to the nine throws listed here, he had a tenth target resulting in a 25-yard DPI. 4. Tyler Boyd CIN 10 10 122 12.2 0 50 SF Boyd had six first downs, including all four of his third-down targets. 5. Kenny Golladay DET 8 10 117 14.6 1 48 LAC Seven of Golladay's targets produced first downs, including a 27-yard gain and a 31-yard touchdown.

 Worst Wide Receiver or Tight End by DYAR Rk Player Team Rec Att Yds Avg TD Total DYAR Opp 1. DeVante Parker MIA 0 7 0 0.0 0 -51 NE While the Dolphins are scouting Tagovailoa, they might notice that DeVonta Smith burned the Gamecocks for eight catches, 136 yards, and two scores. I mean, they've got a lot of draft picks.

## Comments

51 comments, Last at 19 Sep 2019, 3:37pm

### 1The intro reminds me of…

The intro reminds me of something Football Reference wrote years ago about "QB Personality types" splitting quarterbacks into groups based on Meyers Briggs. It was not dissimilar to this and I noticed some of the same things such as former teammates being similar to each other etc.

### 2New Rule Proposed

Each week must have one DeVante/Davante in the top performer and bottom performer table. Just for balance. I think there are a a couple non-WRs in the league who might qualify as well. It's like Michael was 40 years ago.

Also, love the conceptual QB style graphs.I suspect coaches/GMs make similar, but far less technical and specific judgments when looking for their signal callers. They have a "type of guy" they see running their offense. This helps quantify things a bit in that regard. Also, I always thought of Rodgers as quite the slinger, albeit one with an absurdly low INT rate. Maybe those two concepts are mutually exclusive.

### 6Rodgers has and still throws…

Rodgers has and still throws deep a lot, he's just always been the type to hang in to wait for a receiver to break into more space or cross a defender, or to look further downfield than most do while on the move.

Whereas the gunslinger it's easiest for Packers fans like me to compare him to, Favre, would attempt to vaporize his targets on slant and post routes down the middle of the field, all the more mustard on the ball if the window was small or only open for a brief second (or, hell, not at all). He was also good at buying time to throw deeper, but at his MVP peak what set him apart was how he could hit receivers more within the 'regular' timing of the play who functionally weren't open for other quarterbacks.

### 16Favre was Ben before Ben

Favre's ability to shed tacklers and keep a play alive was a legit skill the first part of his career.

### 51Favre's TD passes to offset his INT's

One other thing-Favre's TD to INT differential is so high that it cannot be overlooked. The best thing a QB can do to offset INT's is to throw a lot of TD passes. Favre did just that. Only a handful of guys were better at it-Brady, Peyton, and Brees that I know of.

### 3The QB style graph is…

The QB style graph is interesting. Joey Harrington being an outlier all the way to the right checks out with the way I remember him. He didn't take a huge amount of sacks or throw a ton of interceptions, but watching him try to move the ball down the field was like watching a golden retriever puppy try to climb up a greased ramp.

### 4Odell Beckham Jr.

Did OBJ not make the Top Five? He did catch only 6 of 10 targets (plus one catch that went for no gain), which is I assume is the reason he's missing here. But it's kind of unfair to hold the missed targets against him, since Mayfield will throw him the ball whether he's open or not.

### 19I would also guess the 89…

I would also guess the 89 yard TD looks more impressive by traditional statistics than by DYAR.  Not that it isn’t great, but this play doesn’t entirely make your night by DYAR (and I think this is probably the right evaluate a game in general).

### 5Love the kind of analysis in…

Love the kind of analysis in your intro. It's the best thing about FO (except maybe the best comment threads).

I read this as an impressionistic analysis, an experimental result rather than a finished product.. Plotting the centroid from three different scales certainly loses a lot of information. We can't glean much from anyone near the center of the plot. But it's concise, and gives us good grist for the discussion mill.

If we wanted to look more seriously at this via graphs, we'd need three X-Y scatter plots (ints vs. sacks, ints vs. Y/A, sacks vs. Y/A) or a true three-dimensional plot of all three dimiensions. We'd also scale based on relative value or on standard deviations rather than the ranges form extreme to extreme. The graphs would be technically more useful, but at the expense of being too much to digest in one eye-full. This centroid plot is fine for discussion.

More advanced plots would show us, for example, how P. Manning was similar to late-career Favre in that his one (relative) weakness was interceptions, but was much better even in that dimension than the old Favre. Or that late-career Bledsoe was similarly balanced to Brady, but nowhere near as good. Ryan and Cassel would not look so similar.

### 21It sounds like you could do…

It sounds like you could do a conventional triangle plot from the raw data.

### 7"28-of-36 for 444 yards with…

"28-of-36 for 444 yards with five touchdowns and no interceptions. No, those aren't Rosen's splits in any category. That's what Alabama quarterback Tua Tagovailoa did against South Carolina. Just sayin'. "

Rosen lit up mediocre SEC teams while in college, too. Just sayin'.

### 9What's happened with Rosen…

What's happened with Rosen and Darnold this year is so dispiriting.  I could stand this season if either of them showed some promise, but Rosen's season was doomed before it started (and I wonder if Tua would have a chance in hell in Miami behind that line, throwing to those receivers), and the Jets season has just tanked in an absurd manner.  I was so wrong to  think Gase would implement a competent offense.    At least Lamar Jackson has looked good.

### 10Was kit Testaverde who…

Was it Testaverde who commented that when he went to Tampa from the University of Miami, not only was the opposition much better than what he faced in college, his receivers were actually worse?

That might happen to Tua, too. Also, Tua has badly struggled against any defense that approaches the talent of his own offense. He doesn't project well to the NFL.

### 12Testaverde's receivers in…

Testaverde's receivers in his Heisman year included not just Michael Irvin but Brett Perriman and Brian Blades, both of whom went on to catch more than 500 passes in the NFL. So yeah, in 1987, that was probably better than any corps of NFL receivers east of San Francisco.

### 14Mark Carrier was a pretty…

Mark Carrier was a pretty good receiver who made a Pro Bowl (back in the days when that actually meant something), and went on to have a fine career.  But yeah, he was never going to live up to Irvin as a WR1.  And the other receivers in Tampa during Vinny's career were pretty slim pickings.

### 29Just for the record, Allen…

Just for the record, Allen hasn't exactly sucked this year either. Jury's still out, but at least I feel like Buffalo picked the right Josh.

### 13Did he though? Rosen threw…

Did he though? Rosen threw for 5+ TDs... once, against the Hawaii "defense". He threw for 400+ yards... 6 times, not bad (although he lost 4 of those games, threw interceptions in 5 of them, and was below 10 YPA in all 6). Tua's passing efficiency rating against SC was 227; Rosen had a passing efficiency rating of 200+... twice (technically) - that Hawaii game plus a Cal game where he was benched at halftime after hurting his shoulder on a sack.

### 8Can you include the axes on…

Can you include the axes on that scatterplot?

### 11More of a condensed radar chart

From Vince's description, it sounds like there aren't X and Y axes per se. It's more like a Radar Chart, except instead of plotting a triangle for each QB, he plots the centroid of the triangle.

He could show the three lines for the three dimensions. A QB who was equally bad at sacks and interceptions would have his point show up on the Y/A line

### 20I was trying to figure out…

In reply to by The Powers That Be

I was trying to figure out whether he was using color as a pseudo z-axis on an x-y scatterplot.

### 15I'd be curious to see a Fitz…

I'd be curious to see a Fitz-less second chart to get a less skewed sense of the rest of the field currently playing.

### 22"Bystanders"

Love the look into the data, but don't go into marketing Vince! "Bystanders" seems like it doesn't have any of the connotations you are looking for here. Unfortunately, I found it tough to come up with ones that don't involve allusions to firearms (things like snipers or marksmen - people who have the patience to wait for the good shot and take it) spring to mind - but I'm sure others can probably do better...

...but those guys are *not* bystanders.

### 23The difficulty is that this…

In reply to by RoninX

The difficulty is that this category captures two very different types of QBs.

1. The statues who hang in the pocket while the defense crashes against them, like the prow of a ship in heavy waves.

2. Scramblers.

Maybe the Fashionably Late?

### 38Statues and Scramblers is…

Statues and Scramblers is great.  Could distinguish based on their running yards.  Adds another element to the analysis, though, outside of the current triangle.

Great analysis BTW!

### 24I was going to say the same…

In reply to by RoninX

I was going to say the same thing.  Bystanders implies standing (it's in the name!), which is almost the exact opposite of what most these QBs do.  They run; they scramble; they move.

My proposal for what they should be called: Hanger-ons

### 37Priceless.

In reply to by Aaron Brooks G…

Priceless.

### 25Waiters? Ballholders?…

In reply to by RoninX

Waiters? Ballholders? Patients? Time-biders? Longtimers?

### 27Ballholders because Bevis…

Ballholders because Bevis and Butthead would like it

### 26Ball hogs?

In reply to by RoninX

Ball hogs?

Masochists?

High Pain Drifters?

### 30High pain drifters is just…

In reply to by nat

High pain drifters is just about perfect.

### 35Fram

In reply to by nat

For those old enough to remember the old Fram oil filter commercials, "Pay him later."

### 28Middle

I would say that most that are collected in the middle should be colored grey, as they are the most "whatever fits the situation" style over a hard single style. Ex: having a QB at 5, 7, 7 ranks should not be colored 'red', because he's basically equally all 3, though with the radar graph I guess it would be more a QB that is < x distance from the center would be how it's delineated.

You'd also have blends like K. Collins would be brown and Brunell would be purple, as they seem equally in 2 categories. I guess for me that's, 3 single zones, 1 all-in-one zone, and 3 'corridor' dual zones.

### 31Feedback

Did OBJ not make the Top Five? He did catch only 6 of 10 targets (plus one catch that went for no gain), which is I assume is the reason he's missing here. But it's kind of unfair to hold the missed targets against him, since Mayfield will throw him the ball whether he's open or not.

Well, we listed the top five receivers, and Beckham was not among them, so no, he did not make the top five. He actually finished 18th. You note that he had four incompletions; the top five had a total of five, and none of them had more than two. You can choose to not hold those incompletions against Beckham, but that's not how DVOA or DYAR work -- we're measuring what happened when Team A threw the ball to Player B. Yes, that means all sorts of outside context can affect the numbers. They are not QB-independent or coaching-independent. More info is available on this page:

https://www.footballoutsiders.com/info/methods

It sounds like you could do a conventional triangle plot from the raw data.

It sounds like he's doing something analogous to a ternary phase diagram and is kludging up that kind of plot. (https://en.wikipedia.org/wiki/Ternary_plot)

Thanks for the tips, I'll look into those. For those interested, I wrote a Twitter thread explaining more about how I made the sausage here:

https://twitter.com/FO_VVerhei/status/1174048527929856000

Love the look into the data, but don't go into marketing Vince! "Bystanders" seems like it doesn't have any of the connotations you are looking for here. Unfortunately, I found it tough to come up with ones that don't involve allusions to firearms (things like snipers or marksmen - people who have the patience to wait for the good shot and take it) spring to mind - but I'm sure others can probably do better...

My proposal for what they should be called: Hanger-ons

Waiters? Ballholders? Patients? Time-biders? Longtimers?

Ball hogs?

The one universal bit of feedback I've gotten on this is that everyone hates the name "Bystanders." Fair enough! I am open to suggestions! "Ball holders" does accurately describe what they're technically doing, but no, I'm not using that. I think "Snipers" and "Ball hogs" are my favorites so far.

You'd also have blends like K. Collins would be brown and Brunell would be purple, as they seem equally in 2 categories. I guess for me that's, 3 single zones, 1 all-in-one zone, and 3 'corridor' dual zones.

Sure, there's all kinds of more advanced things you could do with color. The best would probably be to just gradient the whole field with extreme colors in the corners that blend to gray in the middle.

More advanced plots would show us, for example, how P. Manning was similar to late-career Favre in that his one (relative) weakness was interceptions, but was much better even in that dimension than the old Favre. Or that late-career Bledsoe was similarly balanced to Brady, but nowhere near as good. Ryan and Cassel would not look so similar.

The one simple thing I could have is just change the size of the markers, so good QBs get big dots and bad QBs get small dots. Will keep that in mind.

Rosen lit up mediocre SEC teams while in college, too. Just sayin'.

Look, it's entirely possible that Tua Tagovailoa will be a failure as an NFL quarterback, but if you're arguing that Rosen's stats at UCLA are in any way comparable to Tagovailoa's at Alabama, then … no.

### 32The twitter link explains it…

In reply to by Vincent Verhei

The twitter link explains it well.

Your idea of changing the size of plot points to differentiate QBs who are good at everything from ones who are bad at everything is a great idea.

### 41Well, we listed the top five…

In reply to by Vincent Verhei

Well, we listed the top five receivers, and Beckham was not among them, so no, he did not make the top five. He actually finished 18th. You note that he had four incompletions; the top five had a total of five, and none of them had more than two. You can choose to not hold those incompletions against Beckham, but that's not how DVOA or DYAR work -- we're measuring what happened when Team A threw the ball to Player B. Yes, that means all sorts of outside context can affect the numbers. They are not QB-independent or coaching-independent. More info is available on this page:

https://www.footballoutsiders.com/info/methods

Thanks! Hey, if you promise not to be so condescending next time I ask a question, I will promise to include all sorts of disclaimers about how I realize that a player's DYAR is actually measuring the team's performance on the plays in which a certain player is targeted, how I understand that if a player is not listed in the Top Five, that means he didn't actually make the Top Five, and carefully phrase my opinions so that it's clear that they should be held separate from a player's objective assessment per DVOA.

### 43Cut him some slack. Although…

Cut him some slack.

Although he is answering a question that you asked, he's answering it for a the whole audience, including people showing up here for the first time. That means he will at times over-explain something in order to cover an important point that newbies might otherwise miss.

### 44The problem isn't the…

The problem isn't the overexplaining,it's the obvious condescension.

### 46Sorry about that. Looking…

Sorry about that. Looking back now, it comes off a lot harsher than I meant it.

### 45Ball hogs gets my vote!

In reply to by Vincent Verhei

"Ball hogs", I vote for that one.  Applies equally well to Scramblers and Statues, and doesn't require an additional stat to differentiate them.

### 33For the record, Brett Favre…

Since Brett Favre spent most of his career in a Packer uniform, he doesn't really group together well with 'other Jets quarterbacks'. He really was all by himself.

### 36Flacco

I could have sworn Flacco's int was tipped at the line of scrimmage. I wonder if that adjusts his numbers.

### 39Flacco was sharp but...

In reply to by tunesmith

Could have had 3 more INTs.   Floyd dropped one that hit him in the hands.   Eddie Jax also should have had at least one more, if not two.

I don’t see Flacco much as an NFC guy but he was better than his FO reputation (at least for this game).

### 40Hasn't been Falk's Experience

"You'll recall that offenses start with 10 yards to go and then, theoretically, that number gets smaller."

LOL. It might not have been as bad as it was for Mr. Bridgewater, but there were stretches where Luke Falk must have wondered whether that was true at the pro level.

### 42Yeah, I'm honestly pretty…

Yeah, I'm honestly pretty surprised that Falk wasn't "leading" in that category.  Granted it may have been Siemian that truly had the worst of it on Monday.

Speaking of Siemian, I know he didn't have enough snaps to qualify for the board this week, but I'm curious, how bad did he do per DYAR?

### 483/6, 3 yards, 2 sacks, 17…

I was also wondering about the second greatest passer in Northwestern history.

3/6 for 3 yards; 2 sacks for 17 yards; no TDs, picks or first downs.

I'd guess somewhere around -50?

Serious question: does the QB get any sort of credit for roughing the passer calls (of which he drew two), like QBs & WRs do for DPI?

Edit: He appears in the standings for the QB tables. -67 DYAR, -165.5% DVOA, 1.4 QBR, +100 Grittiness rating for walking off the field with a mutilated ankle. Otto Graham, looks like your place atop the Wildcat QB standings is secure!

### 49No, QBs do not get credit…

No, QBs do not get credit for drawing roughing the passer fouls. I'd suspect they are too rare and too random to do any meaningful analysis -- Philadelphia drew the most of any team last year with 10. Tampa Bay drew 9, and Chicago and San Francisco had 8 each. When only four teams can draw a foul in even half their games, there's probably no predictive value there.

### 50That's what I figured…

That's what I figured. Thanks for the explanation.

I thought certain defenses (e.g. whoever employs Clay Matthews) might consistently generate more RTPs than others, but that doesn't appear to be the case. Every team had between 1 and 6, with a five-way tie at the top, (including the Packers).

### 47Yay, the guy the executive of the year traded up for...

...improved to slightly below replacement-level in week 2. By the end of this, his third season as a pro, maybe he'll have positive DYAR and Chicago can throw a parade to celebrate.

PREMIUM STATS & TOOLS
DVOA DATABASE: Exclusive Access
• Historical Lookups
• Single Game DVOA by
• Stat Splits by
• Defense
NFL DVOA Database
• Overall DVOA
• Positions
• Line Stats
NCAA Football Stats
In-Season
Fantasy
NCAA
Offseason