Passer Rating vs. Total QBR (Matt vs. Josh)

[downunderbill]

Eyesight test says he is getting better as the season goes on. Anyone can find stats to justify their opinion.

 

He is the Bills offence so as he goes the game goes. Considering the talent around him he's doing well for such a project QB 

 

[BuffaloHokie13]

If I look at the individual week (ignoring tonight's participants, of course) I've got Josh Allen at 80.0 (7th best) and Matt Stafford at 70.7 (11th best). That mostly has to due with both of them having something in the TD column (11 QBs didn't) and neither of them having something in the Turnover column (17 QBs did).

 

Maybe I'm wasting my time by having my own formula, but I enjoy the time I spend doing it and analyzing the results.

[Alphadawg7]

Just now, dave mcbride said:

It's kinda like wins for starting pitchers - kinda useless, but at the same time, if a guy is winning close to 20 games year and year out, then you can assume he's probably pretty good.

 

Yes totally agree.  When I say worthless, really just that the stat is totally unreliable on its own.  It just lacking a significant part of information and can be heavily misleading.  A guy like Trent Edwards is a good example of what you just said.  Some wins against bad teams and a solid QB rating due to check downs were all fools gold and didn't tell the whole story.  

 

Doesn't mean its always wrong, just means it cant be trusted as a singular analytic tool. 

[dubs]

1 hour ago, JESSEFEFFER said:

 

Those two numbers in context say that Matt Stafford had a much more positive effect on the Lions behalf than Josh did for the Bills.  That's a statement that is unsupported by the game I watched.   

 

Yes, I agree. 

 

Im saying I’m curious as to how this gets calculated. In a game the Bills won, both QBs threw for a TD, neither had an INT and Josh ran for a TD, I don’t understand the disparity in the two ratings. 

[hemma]

I've always hated ESPN's QBR.  

 

Put two football guys in different rooms and ask each to quantify their feelings should be left to professionals like Morris Albert.

 

[Alphadawg7]

32 minutes ago, dubs said:

 

Yes, I agree. 

 

Im saying I’m curious as to how this gets calculated. In a game the Bills won, both QBs threw for a TD, neither had an INT and Josh ran for a TD, I don’t understand the disparity in the two ratings. 

 

I believe the disparity comes from how heavy comp % is weighted in determining the rating.  Take TT for instance, plenty of times he had a strong QBR, but if you watched the game he didn't feel that impactful.  Sure his legs production helped, but so did his tendency to not take risks and only make safe throws which minimized turnovers and incomplete passes.  

 

Josh threw for 50% last night...but did he really is the question.  And the answer is no, he had several drops first of all.  Second, and most importantly, guys like Zay were dominated, and I mean dominated, by their defender all game (in this case by Slay).  There are multiple passes that won't be counted as "drops" because Zay got dominated so badly on the play he didn't even get a hand on the ball when most starting WR's not only make a play on the ball but also make the catch.  He was so often flailing out of control and begging refs for help.  He was manhandled as bad as I have seen, and I dont mean from a PI sense, I mean Slay just ate his lunch then punched him in the mouth.  

 

But the stat sheet still says 50% no matter what, and many of those things just aren't going to show up in a rating system.  Thats the problem all NFL "rating" systems have, they lack context.  Doesn't mean they are always right or always wrong, means no matter what system you use its unreliable.  Football is so much more of a team sport than any other sport out there.  Individual stats are completely dependent by those around you at every position.  Baseball and Basketball, you dont have that nearly as much.  One guy can excel in individual stats while his team isnt good.  Its pretty hard to run behind a bad OL, be a DB if the front 7 cant get any pressure, be a WR with a QB who isnt good, throw to WR's you dont get open and cant catch, return kicks if the guys dont open lanes, etc etc.  

 

Like you basically said, anyone who watched the game though knows that it was his best game as a passer so far.  Accurate and well timed throws all over this game.  And he focused on throwing more than running.

Edited December 17, 2018 by Alphadawg7

[dubs]

1 minute ago, Alphadawg7 said:

 

I believe the disparity comes from how heavy comp % is weighted in determining the rating.  Take TT for instance, plenty of times he had a strong QBR, but if you watched the game he didn't feel that impactful.  Sure his legs production helped, but so did his tendency to not take risks and only make safe throws which minimized turnovers and incomplete passes.  

 

Josh threw for 50% last night...but did he really is the question.  And the answer is no, he had several drops first of all.  Second, and most importantly, guys like Zay were dominated, and I mean dominated, by their defender all game (in this case by Slay).  There are multiple passes that won't be counted as "drops" because Zay got dominated so badly on the play he didn't even get a hand on the ball when most starting WR's not only make a play on the ball but also make the catch.  He was so often flailing out of control and begging refs for help.  He was manhandled as bad as I have seen, and I dont mean from a PI sense, I mean Slay just ate his lunch then punched him in the mouth.  

 

But the stat sheet still says 50% no matter what, and many of those things just aren't going to show up in a rating system.  Thats the problem all NFL "rating" systems have, they lack context.  Doesn't mean they are always right or always wrong, means no matter what system you use its unreliable.  Football is so much more of a team sport than any other sport out there.  Individual stats are completely dependent by those around you at every position.  Baseball and Basketball, you dont have that nearly as much.  One guy can excel in individual stats while his team isnt good.  Its pretty hard to run behind a bad OL, be a DB if the front 7 cant get any pressure, be a WR with a QB who isnt good, throw to WR's you dont get open and cant catch, return kicks if the guys dont open lanes, etc etc.  

 

Anyone who watched the game though knows that it was his best game as a passer so far.  

 

Good post. Thanks. Agree with all you said especially about joshs game. Really liking the kid and feel good about the Bills future with him. 

[JESSEFEFFER]

9 minutes ago, dubs said:

 

Good post. Thanks. Agree with all you said especially about joshs game. Really liking the kid and feel good about the Bills future with him. 

 

11 minutes ago, Alphadawg7 said:

 

I believe the disparity comes from how heavy comp % is weighted in determining the rating.  Take TT for instance, plenty of times he had a strong QBR, but if you watched the game he didn't feel that impactful.  Sure his legs production helped, but so did his tendency to not take risks and only make safe throws which minimized turnovers and incomplete passes.  

 

Josh threw for 50% last night...but did he really is the question.  And the answer is no, he had several drops first of all.  Second, and most importantly, guys like Zay were dominated, and I mean dominated, by their defender all game (in this case by Slay).  There are multiple passes that won't be counted as "drops" because Zay got dominated so badly on the play he didn't even get a hand on the ball when most starting WR's not only make a play on the ball but also make the catch.  He was so often flailing out of control and begging refs for help.  He was manhandled as bad as I have seen, and I dont mean from a PI sense, I mean Slay just ate his lunch then punched him in the mouth.  

 

But the stat sheet still says 50% no matter what, and many of those things just aren't going to show up in a rating system.  Thats the problem all NFL "rating" systems have, they lack context.  Doesn't mean they are always right or always wrong, means no matter what system you use its unreliable.  Football is so much more of a team sport than any other sport out there.  Individual stats are completely dependent by those around you at every position.  Baseball and Basketball, you dont have that nearly as much.  One guy can excel in individual stats while his team isnt good.  Its pretty hard to run behind a bad OL, be a DB if the front 7 cant get any pressure, be a WR with a QB who isnt good, throw to WR's you dont get open and cant catch, return kicks if the guys dont open lanes, etc etc.  

 

Like you basically said, anyone who watched the game though knows that it was his best game as a passer so far.  Accurate and well timed throws all over this game.  And he focused on throwing more than running.

Supposedly this is true only of the NFL's passer rating formula.  Supposedly, ESPN's Total QBR looks at every play in its comparable context and has some regard for clutch play.  Maybe it's the result of a meaningless fumble on a 4th down play that did not gain enough for a first down, one bad sack (really poor block by Mills, I think Josh was rolling to avoid A but got sacked blindside by B who Mills didn't even slow down) and an almost interception?  I suppose those were his three worst plays.

Edited December 17, 2018 by JESSEFEFFER

[K-9]

Someone needs to explain to me how two QBs with nearly identical stat lines can have such disparate QBR ratings. Is it because one took a single sack? Is it because of one inconsequential fumble? Is it because one had a better completion percentage in the first half but didn’t do squat with his final seven possessions? Why isn’t there a consideration for the QB who led his team to a win in the 4th quarter while the other couldn’t? 

 

[Alphadawg7]

11 minutes ago, JESSEFEFFER said:

 

Supposedly this is true only of the NFL's passer rating formula.  Supposedly, ESPN's Total QBR looks at every play in its comparable context.  Maybe it's the result of a meaningless fumble on a 4th down play that did not gain enough for a first down, one bad sack (really poor block by Mills, I think Josh was rolling to avoid A but got sacked blindside by B who Mills didn't even slow down) and an almost interception?

 

Well its true of most ratings and even stats in football.  And yeah, I am sure those contribute, but I think the gap disparity is more heavily rooted in comp % pertaining to Stafford and Josh.  

 

But all stats are fools gold in the NFL...all of them.  Doesn't mean they are always wrong, means they are completely unreliable as stand alone analytic tools.  I used an example earlier...at the end of Gurleys second season, was his low YPC a fair metric to determine his ability?  No, clearly.  His low YPC was a direct result of bad coaching, terrible OL play, bad WR's, and struggling QB's.  He saw stacked boxes all season and the offense stunk.  

 

Or how about Trent Edwards as another example I used.  Here is a guy who had favorable ratings when we were 4-0.  I got flamed non stop because I said after being 4-0 that we would miss the playoffs if Trent kept playing like he was.  People mocked and ranted at me citing his "ratings" and comp % as if they never watched the games.  Missing was the fact we beat the 3 worst teams that year in the NFL and another bad team...and we barely won those games mostly where defensive and special teams TD's were needed to win 3 of them. Missing was the fact Trent didn't really have any strong games, mediocre at best and was constantly checking down.  Ultimately, we missed the playoffs and those "stats" and "metrics" were false positives on Trent as predicted by a small handful of us.

 

Obviously there are a million examples of how a rating or stat gave a false positive conclusion on a player (good or bad).  I certainly feel QBR is by far a better metric than QB Rating, however, it still lacks so much context that is impossible to quantify in a number.  Like the 2 or 3 great passes Allen threw to Zay that won't get logged as "drops" that Zay should have been able to get to but couldn't because hes not good enough and Slay man handled him.  It doesn't factor in talent level, route running, etc etc.  

 

There just isnt a rating that can ever analyze enough data to be truly relied on and always accurate.  I do feel that QBR is about as good as you can get for a QB and most often pretty accurate barometer, just not fully reliable.  But because it can never be perfect, there will be anomalies where guys get a bad score that isnt fair and guys get a good score that isnt deserved.  

Edited December 17, 2018 by Alphadawg7

[chris heff]

73% of all statistics are made up

[Bing Bong]

I saw Stafford's stat line and it was pretty great good. What like 22/29 and a TD? It was more coaching since they never moved the ball in the second half between bad screens, sacks, and runs all for negative plays. That game was ANEMIC. Slow as mollasses. But don't be shocked that Stafford gets a better rating. He's Stat Padford. And he's not a bad quarterback folks.

 

But he did lose, as he is want to do.

 

The game came down to a missed extra point AND field goal. It was a crappy game. As many neutral fans at the bar were very eager to point out to me.

Edited December 17, 2018 by BarkleyForGOATBackupPT5P

[VW82]

5 hours ago, row_33 said:

 

stats are only worth 25-33% at most in a realistic assessment

 

 

 

I think in any kind of analysis you need quantitative and qualitative measurement, and that doesn't mean just one stat plus one non-numerical observation. I think the statistical measurement is worth more than 1/3. 

 

Statistical case: Total QBR of 19.6 (27th), Traditional QB rating of 89.3 (16th), 

Observations: 1. Despite having lots of time it sure looked like guys were struggling to get open. 2. Allen had a number of throwaways and a couple drops that killed drives. 3. Lions did a good job taking away his runs. 4. we had a lot of injuries to key personnel on offense throughout the game. 5. we only scored 14 points at home. 

 

Conclusion: Allen was maybe a little better than his stats suggest, and good enough to win, but when you look at all the available evidence he wasn't incredible or anything. We only scored 14 points!

Edited December 17, 2018 by VW82

[Reks Ryan]

5 hours ago, dave mcbride said:

Team passer rating differential is arguably the most predictive stat in all of major team sports when it comes to team success and failure, and the evidence is fairly overwhelming. I therefore disagree with you, albeit with a caveat: differential isn't the same as an individual qb's passer rating, and unlike differential, sack yards taken aren't factored into individual qb rating. 

This is very true.  Passer rating differential is one of the most reliable stats there is to predict wins and losses.  Here are the top 8 and bottom for this year.   All of the top 8 are in the playoffs or in good position to make it.   And all of the bottom 8, are out or have their playoff chances hanging by a thread.

 

Unfortunately the Bills are 2nd worst  because they're dead last offensively.  They're top 7 Defensively.   It makes sense when you consider the early season blowouts.   I have been impressed with Allen's overall game the last month.  But he will need to significantly  improve efficiency for long term success.

 

2018 Through yesterday.  Passer Rating Differential

Top 8

LA Chargers       +22.1

Chiefs                 +22.0

Bears                   +21.3

Saints                  +19.5

Seahawks           +16.6

Vikings                +13.8

Texans                +10.2

Patriots               +9.4

 

Bottom 8           

Cardinals            -28.5

Buffalo Bills       -26.3

NY Jets                -20.4

Buccaneers        -19.1

Lions                   -16.8

SF 49ers              -13.5

Bengals               -12.4

Redskins             -11.1

[The Frankish Reich]

Baseball is easy: (1) it's a game of individual achievement in a team setting; (2) there's now general agreement on how much a certain component (example: a home run, a stolen base, a caught stealing) contributes to a win. So while there are different formulations of WAR (wins above replacement player), the general concept is agreed upon ... now, pretty much by anyone who counts: GMs, player agents, even Hall of Fame voters.

Football is tough: (1) it's a game in which it's exceedingly difficult to isolate individual performance from a team setting; (2) there's not a ton of agreement on what components create success or failure. But understanding that, passer rating is pretty good. It passes the eye test -- sure, there are guys who seem able to "game" the rating system (probably not deliberately) by avoiding bad outcomes. Tyrod Taylor, step right up: you minimize interceptions by refusing to throw into tight coverage; you minimize incompletions by throwing short passes. That gets you into the median range for starting QBs, typically around the 15th-20th highest rated passer. But here's what it doesn't do: it doesn't get you into the elite category. This year, the top 15 QBs on passer rating are:

Brees

Mahomes

Rivers

Wilson

Ryan

Watson

Wentz

Tannehill (he has that Tyrod Cheat Code too)

Fitzpatrick (don't laugh ...)

Cousins

Luck

Goff

Brady

Rodgers

Newton

And if you take a weighted average over multiple years, the quirky ones (Fitzy) will drop off that list, and you'll wind up with ... pretty much a consensus ranking of the Top 10 or 15 QBs in football.  That's the sign of a good stat. It helps because it allows us to make more subtle gradations, etc., but that's what it is -- a ranking, and it's a good one. It's not the be all and end all, but it's pretty damn accurate despite being kind of a back-of-the-envelope thing the NFL came up with 35 years ago or so.

I use it alongside ANY/A (adjusted net yards per attempt - you'll also see that on profootball reference), and - pay attention Allen fans! - espn's Total QBR when we're talking about the relative value of a guy who creates a lot of value by running. It's kind of a seat of the pants amalgamation of all 3 for me.

So to say passer rating is useless?  That's a mistake my friends. It's very useful, but it ain't everything. If you've got something better other than "this is my personal ranking, let's order a pitcher and argue about it," then bring it on ...