NFL Stats

padraig_f

I thought it might be useful to create a thread about the discussion of NFL stats.

Stats can be a contentious subject, so I just want to say I think no stat is perfect, and will never replace game watching. At the same time, no-one has time to watch all the games, so stats can be a useful tool for summarising games and providing some objective facts.

Stats are there to be challenged, no number can adequately capture the complexity of an NFL game, but they can still provide insight and do certain things better than people watching the game, so there's a place for both.

j8wk2feszrnpao

Before the season started someone had the opinion that Brady was was finished as an elite/top5 QB, and then went into stats mode to 'prove' it.
Doubt that opinion is still voiced too loudly.

IMO without in-depth analysis and actually watching the games; stats can used and skewed by some to 'prove' their opinion.

eagle eye

Best stats out there are the football outsiders stats.

padraig_f

I calculate yards-per-play efficiency each week to highlight good and bad performances.

Here are the numbers for week 9:

team_id 	overall		offense			defense (lower is better)

MIA 		2.03 		0.25			-1.77
DEN  		1.34 		0.83 			-0.51
PIT  		1.11 		0.49 			-0.61
PHI  		1.01 		0.62 			-0.39
ARI  		0.85 		-0.86 			-1.72
TB  		0.72 		0.37 			-0.35
WSH  		0.66 		0.58 			-0.07
SF  		0.47 		-1.35 			-1.82
OAK  		0.18 		-1.41 			-1.60
SEA  		0.15 		-1.18 			-1.33
JAX  		0.08 		0.75 			0.66
NE  		0.05 		0.12 			0.06
NO  		-0.09 		-0.93 			-0.84
MIN  		-0.16 		-0.51 			-0.34
NYG  		-0.18 		-0.05 			0.12
NYJ  		-0.21 		-0.12 			0.08
CIN  		-0.37 		0.69 			1.06
KC  		-0.53 		0.11 			0.65
CAR  		-0.54 		-1.81 			-1.26
BAL  		-0.66 		-0.91 			-0.24
STL  		-0.84 		-1.61 			-0.76
IND  		-0.84 		-0.03 			0.81
HOU  		-0.93 		-0.18 			0.75
SD  		-1.37 		-1.05 			0.32
DAL  		-1.38 		-1.30 			0.07
CLE  		-1.60 		-0.76 			0.83

Number 1 is Miami for their 37-0 demolition of San Diego. The numbers are adjusted for strength-of-opponent, so Miami gets extra credit for shutting out a good offense like San Diego. And you see from the numbers, it was done mostly with defense (although I guess the 0 points conceded was another indication of that). In my numbers, Miami has moved ahead of Seattle as the best pass-defense in the league.

Best performance in a loss is Denver, who actually have a higher yards-per-play rating for the game than New England. Possibly Denver's numbers are a bit inflated here by the Patriots going into a prevent-type defense after getting a big lead, but this game was closer than the final score suggested. If you look at the box-score Denver finished the game with higher yards-per-pass numbers and similar rushing efficiency.

Why did New England win? One less turnover, punt return for a TD, missed field goal for Denver, stalled redzone drives for Denver, more 3rd & 4th down conversions for New England. Watching the game I felt it was a game where things went for the Patriots. Not denying them their win, but his is the point of these stats, to highlight some things the score doesn't.

Another good performance in a loss was San Francisco. They had the best defensive performance of the week, unfortunately they also had a very poor offensive performance. They held St. Louis to 3.7 yards-per-play, forced two interceptions, and still lost the game.

At the end of the game, down 13-10, the 49ers had 1st-and-goal from the 2. Likely win and you think they're at least going to overtime. Two passes, one incomplete, one held short. On 3rd-down Kaepernick fumbled the ball on a QB sneak, recovered by St. Louis to secure the win. Kind of a freak sequence in a game the 49ers should've won.

padraig_f

eagle eye wrote: »

Best stats out there are the football outsiders stats.

Yeah they are for team stats (Pro-football Focus probably best for individual stats). They calculate their stats similar to mine (in that we both adjust for strength of opponent), although they have more resources than me to do it on an individual per-play basis. I see they are also raving about the Miami defense this week, who they now rank 2nd-best defense in the league:

Speaking of best performances of the year, how about those Miami Dolphins? I will fully admit that I did not believe in this team at all before the season. They just seemed like an easily ignorable monument to mediocrity, much like their division rivals in Buffalo -- and if one of those teams was going to get its act together, we wrote that it was more likely to be Buffalo. Well, the Bills did get their act together a little bit, but not as much as the Dolphins. Miami was already No. 10 in DVOA before this week but a 37-0 win over San Diego vaults them all the way up to No. 3. The Dolphins also pass Denver and now rank second in defense, the only team behind Detroit. Opponent adjustments are still going to change over the next eight weeks but for now, the Dolphins have the best single game of the year, with 117.8% DVOA for this win.

Billy86

Carried over from the other thread...

Putin wrote: »

So they were easy wins that's all.

They were easy because the Packers had a big lead. The Packers had a big lead because...

Areyouwell wrote: »

Meaningless hocus pocus. I've seen games where Brees throws 3-4 TDs with no Ints, Manning throws 3-4 TDs but has 1 Int. Apart from the one Int, the performances are identical. Yet Manning comes out with a higher quarterback rating for the game . So I wouldn't be losing the run of myself with these kind of nonsense stats, because some of them make no sense at all.

Funny stat to bring up. The second highest TD:INT ratio of all time is Tom Brady with 2.78 TDs to every interception, followed by Manning at 2.30:1, then Steve Young at 2.17:1. These three all time great are separated by just 0.61 TDs to every interception.

The highest is Aaron Rodgers with 3.76 TDs to every INT - basically a full touchdown better. And for most of his prime he has done this with a worse running game than almost everyone in the league, and an offensive line on average worse than most.

Areyouwell wrote: »

Meaningless hocus pocus. I've seen games where Brees throws 3-4 TDs with no Ints, Manning throws 3-4 TDs but has 1 Int. Apart from the one Int, the performances are identical. Yet Manning comes out with a higher quarterback rating for the game . So I wouldn't be losing the run of myself with these kind of nonsense stats, because some of them make no sense at all.

Funny stat to bring up. The second highest TD:INT ratio of all time is Tom Brady with 2.78 TDs to every interception, followed by Manning at 2.30:1, then Steve Young at 2.17:1. These three all time great are separated by just 0.61 TDs to every interception.

The highest is Aaron Rodgers with 3.76 TDs to every TD - basically a full touchdown better.

Billy86

Just to lob in some others as well as well, TDs per start:
Rodgers - 2.17
Manning - 2.08
Brees - 1.96
Brady - 1.90
Luck - 1.76
Rivers - 1.75
Roethlisberger - 1.59

Rodgers is also higher than Marino, Young or Favre here so likely the highest of all time. Marino was the highest of that bunch with 1.90.

Yards per attempt, career:
Rodgers - 8.2
Rivers - 7.9
Roethlisberger - 7.9
Manning - 7.7
Brees - 7.5
Brady - 7.4
Luck - 7.1

Rodgers is 3rd all time here, behind Otto Graham (9.0) and Sid Luckman (8.4) who both played in the 40s and 50s and are in most well read historians top 5's of all time (Graham especially, who is #1 for many).

Pass completion percentage:
Brees - 66.1%
Rodgers - 66.0%
Manning - 65.5%
Rivers - 64.6%
Roethlisberger - 63.6%
Brady - 63.5%
Luck - 58.6%

Brees is first all time here, followed by Rodgers and Chad Pennington in 2nd and Manning in 3rd.

Percentage of passes that ended in TDs:
Rodgers - 6.5%
Manning - 5.9%
Rivers - 5.5%
Brady - 5.5%
Brees - 5.3%
Roethlisberger - 5.1%
Luck - 4.5%

Rodgers is 5th all time here, behind Graham and Luckman again as well as two other 40s/50s QB in Frankie albert and Frank Ryan. Next on the list is Manning, down in 14th.

Percentage of passes interceptions:
Rodgers - 1.7%
Brady - 2.0%
Luck - 2.3%
Rivers - 2.5%
Manning - 2.6%
Brees - 2.6%
Roethlisberger - 2.7%

Rodgers has the best percentage in league history, Brady the second best, and tied Luck in 6th with Mark Brunell and Jeff Garcia.

Now why oh why would only 7 of Aaron Rodgers 63 games

Hazys

You can prove anything with stats.

Stats can only be taken with a grain of salt. Football has too many variables for stats to be anyway effective. I don't think there will ever be a statistical equation that will every be able to accurate predict outcomes of games or used to evaluate players better than the naked eye, a supplement sure but nothing that's it.

Sports like baseball maybe but certainly not football.

Hazys

Funny quote from Roy Keane's book when they started use a new technology to track player performance.

Roy questioned a player's performance in a game, the player replied "well according to the analytics I ran 12 miles", Roy replied "you ran 12 miles because you kept giving the fcuking ball away!"

Areyouwell

Billy86 wrote: »

Carried over from the other thread....

I wasn't addressing you in the last thread. Your post is also quiet confusingly structured. You seem to have quoted the same quote twice and also duplicated your response.

Billy86 wrote: »

They were easy because the Packers had a big lead. The Packers had a big lead because.....

Because what?

Billy86 wrote: »

Funny stat to bring up..

What stat? I was talking about QBR & overall game ratings. I went on to give a specific example between Rodgers & Brady. I was speaking about two recent wonderful performances from Rodgers & Brady. Brady gets a higher QBR and a lower overall rating. Yet he did more of everything compared to Rodgers in his game (TDS, Comps, Yards ect) . I used this to illustrate the hocus pocus of such ratings, the mysterious algorithms they use and why I disregard these kind of stats.

So where did TD:INT ratios come from And what has it got to do with the conversation I had earlier with another poster?

Billy86 wrote: »

The second highest TD:INT ratio of all time is Tom Brady with 2.78 TDs to every interception, followed by Manning at 2.30:1, then Steve Young at 2.17:1. These three all time great are separated by just 0.61 TDs to every interception.

The highest is Aaron Rodgers with 3.76 TDs to every INT - basically a full touchdown better..

I think that there exemplifies perfectly how some stats can be easily manipulated to suit. Are you seriously comparing Rodgers to Manning & Brady? I mean these guys have been NFL starters for the double the length of time & games that Rodgers has. So lets be realistic here and maybe come back to this thread in 8 years time, when Rodgers has a similar number of years & games under his belt. Lets see then what his stats and ratios are.

Billy86 wrote: »

And for most of his prime he has done this with a worse running game than almost everyone in the league, and an offensive line on average worse than most..

Manning, Brady, they've all led pass dominated offenses. They never had dominant unstoppable running games. So what if Rodgers hasn't a great running game. He's not the only one that's had that during his career. The Broncos currently have the 27th ranked RG in the league, The Packers 24th and the Pats 19th. And last year the Packers had the 7th best running game in the league, higher than the Broncos & the Pats.

And when it comes to the O line? Rodgers almost doesn't even need it, he's always been a very good mobile QB. He can just run around, buying time and someone usually gets open. When a QB can improvise like that, the O line certainly won't affect him to the same extent as it would an immobile QB or classic pocket QB.

matthew8

A quarterback who posts incredible stats and doesn't even need an offensive line nor a running game and yet isn't a great.

You keep on banging on about the Rodgers vs. Panthers game too. In fact that's a great example as to why passer rating should be proportional (stop mentioning it with QBR btw, they're measuring different things). Aaron Rodgers and Tom Brady kept going until their opponents were blown out, finished. This happened quicker for Rodgers as the Panthers barely scored. The job was done so he didn't have any reason to keep going. Why would you punish this in the rating? Also, just because a rating system isn't absolute doesn't mean it's useless. Look at all the stock market ratings. Are you guaranteed a profit by following them? Obviously not. Do extremely smart and rich people make use of them as well as other info? Hell yes. What about horse racing handicaps? They should all reach the line at roughly the same time in every race or else the whole system is useless, right? Funnily enough I've never heard a commentator call the system hocus pocus because one horse finished far ahead of the rest.

Areyouwell

matthew8 wrote: »

A quarterback who posts incredible stats and doesn't even need an offensive line nor a running game and yet isn't a great.

Well he had a better running game that Manning & Brady last year. And where did I state that Rodgers wasn't a great QB? I've only pointed out the stupidity of some stats by highlighting a recent anomaly. I never said Rodgers wasn't a great QB.

matthew8 wrote: »

You keep on banging on about the Rodgers vs. Panthers game too.

Dafuq? Are you reading the posts? In fact, the only reason I posted in the first place, was when you seemed to lose the run of yourself a little bit here....

matthew8 wrote: »

Statistically Rodgers has the highest quarterback rating of all time and that's not changing any time soon.

I thought the best way to respond to that kind of boasting, was to highlight how nonsensical some statistics can be with this...

Areyouwell wrote: »

Here's another example of hocus pocus. Against the Panthers Rodgers was 19/22 - 86.4% CR, for 255 yards, 3 TDs, 0 Ints The QBR was 94.3 and his rating was 154.5.

A week later against the Bears Brady was 30/35 - 85.7% CR, for 354 yards, 5 TDs, 0 Ints. His QBR was higher @ 98.8 but the overall rating was lower @ 148.4 . Brady throw 2 more TDs and completes 11 more passes for 101 more yards and the overall rating is lower . This is just the most recent example I could find, to just point out some of the nonsense ratings the NFL often throws out.

You’re the one that brought up ratings with your Rodgers boast. I merely gave a recent example in the above quote, of how much nonsense is involved with such stats. So I simply asked, how does Brady have a higher QBR and yet has a lower overall rating, despite him doing more than Rodgers?

matthew8 wrote: »

In fact that's a great example as to why passer rating should be proportional (stop mentioning it with QBR btw, they're measuring different things).

What? Haven’t I been talking about these two separate ratings all along? And where did I say both measures the same thing? Which again takes me back to my original point. Can you tell me why Brady has a higher QBR rating but a lower overall rating? Yes we all know they measure different things, but how is possible to have a higher QBR, throw more TDs, throw more completions and have a lower overall rating?

matthew8 wrote: »

Aaron Rodgers and Tom Brady kept going until their opponents were blown out, finished. This happened quicker for Rodgers as the Panthers barely scored. The job was done so he didn't have any reason to keep going.

At half time the Packers were leading by 25 point, the Pats were leading by 31 points. The Pats clearly had the bigger lead at half-time. Rodgers didn’t play in the 4th quarter; when he left the field the Packers were up 25 points. Brady played 5 minutes into the 4th quarter; when he left the field the pats were up by 33 points (scored a FG on his last drive). So your assumption that Rodgers finished the game off quicker than Brady is wrong.

matthew8 wrote: »

Also, just because a rating system isn't absolute doesn't mean it's useless.

Yes it is useless, because if it's not fit for purpose and cannot provide 100% accuracy, then it's a complete waste of time. So a system that's not absolute shouldn't' be used then. Yards, Passes, Completions, Ints, Fumbles, these are just some of the stats most fans are interested in. These are stats that are easily & accurately measured. Not some bull**** QBR rating or overall rating that has to be determined by a computerised process, that seems to make no sense & nobody can really explain.

Rodgers




Brady





So when you compare the above stats, I'm just trying to figure out why Rodgers has a higher overall rating in his game, when Brady did more of everything in his?

matthew8 wrote: »

Look at all the stock market ratings. Are you guaranteed a profit by following them? Obviously not. Do extremely smart and rich people make use of them as well as other info? Hell yes. What about horse racing handicaps? They should all reach the line at roughly the same time in every race or else the whole system is useless, right? Funnily enough I've never heard a commentator call the system hocus pocus because one horse finished far ahead of the rest.

Huh? How has any of that had got anything to do with an NFL game? Tbh, I looked at the time you posted that and assumed you had a few in the tank. But let’s use the horse race analogy. A horse wins by a neck or a length, this is something we all can visibly observe with our own eyes. We see it and we don’t need a Jockey Club computer kicking in, that suddenly computates that the horse that came second won.

When two QBs play great football, and one throws a bit more of everything than the other guy. Logic and reason should dictate that he’ll probably have a slightly higher rating. But logic appears to go out the window when the NFL mish mash of algorithms gets to work. And sorry, I’m sounding like a broken record at this stage. But I'm hoping for someone to explain how a player who did more, gets a lower overall rating.

poldebruin

padraig_f wrote: »

I thought it might be useful to create a thread about the discussion of NFL stats.

Great idea Padraig. I love reading about statistics, analysing them - don't create my own. Possibly would have when younger though!

padraig_f wrote: »

Stats can be a contentious subject, so I just want to say I think no stat is perfect, and will never replace game watching.

Couldn't agree more. The biggest challenge imo is comparing across different eras. I'm also growing a bit tired of the " he had nothing around him" argument. These have gotten completely out of hand recently it would seem (on boards.ie at least)

From a personal point of view I knew statistics couldn't tell the whole story, because the team I followed had a great QB that just doesn't (and didn't) measure up statistically, yet when I watched the games I knew he was playing the position at extremely high level and was doing jaw dropping things - John Elway. The two things didn't correspond for me. Now I can see what was driving a lot of that.

padraig_f wrote: »

Stats are there to be challenged, no number can adequately capture the complexity of an NFL game, but they can still provide insight and do certain things better than people watching the game, so there's a place for both.

They give a useful insight, but are no substitute for watching games. They can throw up interesting angles that you never considered before. For instance, did you know that throwing more interceptions, or having a higher interception rate can be a good thing? It's counterintuitive and I had never considered it, but logically makes perfect sense.

When trailing by 2 scores or more late in a game, you would expect to (hope) to see your team and QB be more aggressive in driving the ball down the field. Instead of little dump offs and draw plays, you'd want to see more risks taken in order to get back into the game. Therefore, a good QB (that knows the game situation)should see their int. rate increase in this scenario.

The Aaron Rodgers enigma

poldebruin

Areyouwell wrote: »

Rodgers




Brady





So when you compare the above stats, I'm just trying to figure out why Rodgers has a higher overall rating in his game, when Brady did more of everything in his?

Well the first thing to realise is that the NFL QB Passer Rating method, is outdated and not a particularly accurate model for recognising total QB play.
It disproportionately rewards high completion %. Does not take into account game situation, garbage time yards, score at the time, adjusted to account for strength of opposition, does not recognise rushing yards for QB or sacks taken. So it is flawed. ESPN's QBR tries to address that, and while it makes some stridfes in that area, ~I think it needs to have it's own formulas tweaked, and will likely do so over time. Some of the advanced statistics sites grade out players based on every play and does account for game situations. Advanced Football Analytics use EPA (expected points added) as a good measure of a players value during a game.

Pro Football Focus also grades out players on a per play basis.
Brady

This performance goes onto their hall of fame page.

Rodgers

They make the point that Rodgers faced no pressure and that 3/4 of the yardage came after the receivers caught the ball. This still requires correct decision making and ball placement on the part of the QB though, which does not seem to be taken into account.

These things could not be ascertained by simply looking at the box score, where the performances look very similar.

Looking at the statistics from the above game, the raw statistics are so close and the final ratings reflect that, I don't really see the discrepancy you are mentioning. Arguing over decimal points in a rating that is accepted as being not particularly accurate is a waste of time.

Areyouwell wrote: »

When two QBs play great football, and one throws a bit more of everything than the other guy. Logic and reason should dictate that he’ll probably have a slightly higher rating. But logic appears to go out the window when the NFL mish mash of algorithms gets to work. And sorry, I’m sounding like a broken record at this stage. But I'm hoping for someone to explain how a player who did more, gets a lower overall rating.

I assume you mean Rodgers receiving a higher rating than Brady using the NFL passing rating method.....

If you look at the formula they use you can see why.

Why the NFL passing rating does not work

Too much emphasis is placed on Completion % and Rodgers shaded Brady in this category. Also, it does not differentiate between 3 TDs in 21 passes and 5 in 35 as it calculates TD% but we can clearly use common sense to determine 5 in 35 throws is a better performance and more difficult to do.

Finally, if you want to know why passer rating does not sum up a qbs performance in a nutshell, look up Chad Pennington's career stats. He won the passing title in 2002 despite being a thoroughly average QB.

taidghbaby

munstermagic11 wrote: »

Before the season started someone had the opinion that Brady was was finished as an elite/top5 QB, and then went into stats mode to 'prove' it.
Doubt that opinion is still voiced too loudly.

IMO without in-depth analysis and actually watching the games; stats can used and skewed by some to 'prove' their opinion.

In fairness it's been PFF that have been pushing the Brady is on a down curve opinion, and their stats are about as in depth as you'll find along with Football Outsiders!

Now whether you agree with their scoring of QB's or not is a completely different matter! There's been a lot of "controversy" on their scoring of Luck following the Giants game when they graded him negatively! Now I know Luck didn't play well and got lucky on a number of occasions but to say he his play on the night was more detrimental than beneficial to the team when the blew out the Giants on the road is ridiculous!

matthew8

Areyouwell wrote: »

Well he had a better running game that Manning & Brady last year. And where did I state that Rodgers wasn't a great QB? I've only pointed out the stupidity of some stats by highlighting a recent anomaly. I never said Rodgers wasn't a great QB.



Dafuq? Are you reading the posts? In fact, the only reason I posted in the first place, was when you seemed to lose the run of yourself a little bit here....



I thought the best way to respond to that kind of boasting, was to highlight how nonsensical some statistics can be with this...



You’re the one that brought up ratings with your Rodgers boast. I merely gave a recent example in the above quote, of how much nonsense is involved with such stats. So I simply asked, how does Brady have a higher QBR and yet has a lower overall rating, despite him doing more than Rodgers?



What? Haven’t I been talking about these two separate ratings all along? And where did I say both measures the same thing? Which again takes me back to my original point. Can you tell me why Brady has a higher QBR rating but a lower overall rating? Yes we all know they measure different things, but how is possible to have a higher QBR, throw more TDs, throw more completions and have a lower overall rating?



At half time the Packers were leading by 25 point, the Pats were leading by 31 points. The Pats clearly had the bigger lead at half-time. Rodgers didn’t play in the 4th quarter; when he left the field the Packers were up 25 points. Brady played 5 minutes into the 4th quarter; when he left the field the pats were up by 33 points (scored a FG on his last drive). So your assumption that Rodgers finished the game off quicker than Brady is wrong.


Yes it is useless, because if it's not fit for purpose and cannot provide 100% accuracy, then it's a complete waste of time. So a system that's not absolute shouldn't' be used then. Yards, Passes, Completions, Ints, Fumbles, these are just some of the stats most fans are interested in. These are stats that are easily & accurately measured. Not some bull**** QBR rating or overall rating that has to be determined by a computerised process, that seems to make no sense & nobody can really explain.

Rodgers




Brady





So when you compare the above stats, I'm just trying to figure out why Rodgers has a higher overall rating in his game, when Brady did more of everything in his?





Huh? How has any of that had got anything to do with an NFL game? Tbh, I looked at the time you posted that and assumed you had a few in the tank. But let’s use the horse race analogy. A horse wins by a neck or a length, this is something we all can visibly observe with our own eyes. We see it and we don’t need a Jockey Club computer kicking in, that suddenly computates that the horse that came second won.

When two QBs play great football, and one throws a bit more of everything than the other guy. Logic and reason should dictate that he’ll probably have a slightly higher rating. But logic appears to go out the window when the NFL mish mash of algorithms gets to work. And sorry, I’m sounding like a broken record at this stage. But I'm hoping for someone to explain how a player who did more, gets a lower overall rating.

I'm done. You're taking the piss here. You think I've been on the piss, I think you've stepped out of a time machine from the middle ages. I've said it a million times. The rating system isn't perfect, it's not meant to be perfect, rating systems aren't. It's a proportional to volume summary of other statistics that is not intended to tell the whole story of a particular game. And it's not something you need a huge computer to calculate, you can do it on any decent calculator:

It's also not called the "overall rating". It's called the passer rating. And the only reason I posted it was to save myself the time it would take to post how Rodgers, per start, leads all the statistics apart from yards per attempt, which is lead by people who played the game when it was very different.

eagle eye

Rodgers is a great QB but he has a long career ahead of him unless he gets a bad injury. When he starts to slow down a bit and becomes less elusive we will see exactly where he stands in the GOAT conversation. Right now he is a couple of superbowls away from earning the title.

padraig_f

poldebruin wrote: »

When trailing by 2 scores or more late in a game, you would expect to (hope) to see your team and QB be more aggressive in driving the ball down the field. Instead of little dump offs and draw plays, you'd want to see more risks taken in order to get back into the game. Therefore, a good QB (that knows the game situation)should see their int. rate increase in this scenario.

The Aaron Rodgers enigma

Good article, hadn't noticed that about Rodgers before, but the numbers seem significant...0 wins in 21 games when trailing by 9 or more, compared to Manning's 14 wins in 49 games.

Not knocking Rodgers, I think he's a great QB, but no player is perfect and maybe that's one of his few weaknesses.

That whole phenomenon of interceptions is interesting alright. Alex Smith gets praised for taking care of the ball, but I'd say he hurts his team by not taking enough chances. He's another one that probably could do better with a slightly higher interception rate

matthew8

padraig_f wrote: »

Good article, hadn't noticed that about Rodgers before, but the numbers seem significant...0 wins in 21 games when trailing by 9 or more, compared to Manning's 14 wins in 49 games.

Not knocking Rodgers, I think he's a great QB, but no player is perfect and maybe that's one of his few weaknesses.

That whole phenomenon of interceptions is interesting alright. Alex Smith gets praised for taking care of the ball, but I'd say he hurts his team by not taking enough chances. He's another one that probably could do better with a slightly higher interception rate

Rodgers trailed by 18 when coming back against the Jets. Must be a second half or 4th quarter stat.

Areyouwell

poldebruin wrote: »

Well the first thing to realise is that the NFL QB Passer Rating method, is outdated and not a particularly accurate model for recognising total QB play.
It disproportionately rewards high completion %. Does not take into account game situation, garbage time yards, score at the time, adjusted to account for strength of opposition, does not recognise rushing yards for QB or sacks taken. So it is flawed. ESPN's QBR tries to address that, and while it makes some stridfes in that area, ~I think it needs to have it's own formulas tweaked, and will likely do so over time. Some of the advanced statistics sites grade out players based on every play and does account for game situations. Advanced Football Analytics use EPA (expected points added) as a good measure of a players value during a game.

Pro Football Focus also grades out players on a per play basis.
Brady

This performance goes onto their hall of fame page.

Rodgers

They make the point that Rodgers faced no pressure and that 3/4 of the yardage came after the receivers caught the ball. This still requires correct decision making and ball placement on the part of the QB though, which does not seem to be taken into account.

These things could not be ascertained by simply looking at the box score, where the performances look very similar.

Looking at the statistics from the above game, the raw statistics are so close and the final ratings reflect that, I don't really see the discrepancy you are mentioning. Arguing over decimal points in a rating that is accepted as being not particularly accurate is a waste of time.





I assume you mean Rodgers receiving a higher rating than Brady using the NFL passing rating method.....

If you look at the formula they use you can see why.

Why the NFL passing rating does not work

Too much emphasis is placed on Completion % and Rodgers shaded Brady in this category. Also, it does not differentiate between 3 TDs in 21 passes and 5 in 35 as it calculates TD% but we can clearly use common sense to determine 5 in 35 throws is a better performance and more difficult to do.

Finally, if you want to know why passer rating does not sum up a qbs performance in a nutshell, look up Chad Pennington's career stats. He won the passing title in 2002 despite being a thoroughly average QB.

Thanks poledebruin some very interesting and helpful stuff their, much appreciated.

Areyouwell

matthew8 wrote: »

I'm done. You're taking the piss here. You think I've been on the piss, I think you've stepped out of a time machine from the middle ages. I've said it a million times.

And this coming from the person who went on some gobbledygook tangent about the stock market and horse racing last night? And you accuse me of taking the piss? Please. It's pretty clear from my post, that I'm looking for an answer to what was a recent anomaly in two games. And I also thought it was obvious when I mentioned overall rating, that I was speaking about the overall passer rating. So if it wasn't clear, then I apologise for any confusion.

Vandango

taidghbaby wrote: »

In fairness it's been PFF that have been pushing the Brady is on a down curve opinion, and their stats are about as in depth as you'll find along with Football Outsiders!

Gets to the AFC title game last year with no offense or O line. When you look at last years stats you see a dip. But what the stats on the piece of paper don't tell you is, here is a QB entering his season with his top 6 receivers gone and 94% of passes from 2012 gone. Fecking brilliant season he had last year, especially when you step back and look at the bigger picture & background. And that's the limitations of stats, they never tell you the whole story or the context to how they were achieved.

Billy86

Vandango wrote: »

Gets to the AFC title game last year with no offense or O line. When you look at last years stats you see a dip. But what the stats on the piece of paper don't tell you is, here is a QB entering his season with his top 6 receivers gone and 94% of passes from 2012 gone. Fecking brilliant season he had last year, especially when you step back and look at the bigger picture & background. And that's the limitations of stats, they never tell you the whole story or the context to how they were achieved.

To be fair, a top 10 running game (both in total yards and YPC) does kind of count as part of the offence. Not to say he did anything but great with extremely limited passing resources earlier on in the year especially, but it's still helps to be able to get over 2,000 yards on the ground.

padraig_f

matthew8 wrote: »

Rodgers trailed by 18 when coming back against the Jets. Must be a second half or 4th quarter stat.

Yeah you're right, it's trailing by 9+ in the 2nd half.

Vandango

Billy86 wrote: »

To be fair, a top 10 running game (both in total yards and YPC) does kind of count as part of the offence.

1) Where was I talking about a running game?

2) Who said rushing isn't part of an offense?

Every team has a running game. And how did the running game, help Brady overcome losing his top 6 receivers and 94% of passes last year? And why then did the Pats run the ball less last year, than they did in 2012? In 2012 the Pats had more rushing yards and had 53 more rushing attempts than all of last year.

Billy86

Areyouwell wrote: »

I wasn't addressing you in the last thread. Your post is also quiet confusingly structured. You seem to have quoted the same quote twice and also duplicated your response.

The joys of posting off your phone on the bus back from the airport. And it's a discussion thread - people are free to comment on other people's posts, that's how it works.

Because what?

Because when your quarterback has the most touchdowns per attempt, most yards per attempt, most touchdowns per start, lowest interceptions per attempt, and second highest completion percentage (only 0.1% off tying the #1 guy) in NFL history... you don't tend to need to come back in the fourth quarter much.

I was also getting at fourth quarter comebacks being quite a meaningless stat to judge QBs alone on - for example, Rodgers has 32 regular season losses in games he has started (31 if you discount the Bears game where he got injured after 2 or 3 pass attempts)... in those games he has brought the score back to at least level 20 times in the fourth quarter, only for his defence to give up points right after again in 13 of them (65%), 10 of which were inside the 2 minute warning or OT and 9 of which were within the last 40 seconds or OT. To call Green Bay's defence over Rodgers' time starting depressing would be a massive, massive understatement.

By comparison, Manning has brought his team(s) back to either a tie score or a lead in the fourth quarter 51 times, and only on 11 of those have his defence cost them - 21.6% of the time. Brady has brought the Patriots back to either a tie score or a lead in the fourth quarter 37 times, and on only 6 of those occasions has his defence cost them - 16.2% of the time. They are both massive differences to Rodgers' defence costing the team 65% of the time when he has come back to get the lead or tie the game in the fourth quarter.

What stat? I was talking about QBR & overall game ratings. I went on to give a specific example between Rodgers & Brady. I was speaking about two recent wonderful performances from Rodgers & Brady. Brady gets a higher QBR and a lower overall rating. Yet he did more of everything compared to Rodgers in his game (TDS, Comps, Yards ect) . I used this to illustrate the hocus pocus of such ratings, the mysterious algorithms they use and why I disregard these kind of stats.

So where did TD:INT ratios come from And what has it got to do with the conversation I had earlier with another poster?

You were responding to a post about Rodgers having the highest QBR of all time by mentioning that the rating systems seem to ignore a lot of stats, and you brought the TD:INT ratios into it to prove our point. I was pointing out that Rodgers being #1 all time at YPA, TD per attempt, INT per attempt, TD per start, TD:INT ratio and #2 at completion percentage by just 0.1% shows why he is #1 all time in many QBR systems. The asterisk there has to be the 2004 rule changes, but as for stats being misleading in terms of the pure numbers they most certainly are not with Rodgers being #1 overall on QBR.

I think that there exemplifies perfectly how some stats can be easily manipulated to suit. Are you seriously comparing Rodgers to Manning & Brady? I mean these guys have been NFL starters for the double the length of time & games that Rodgers has. So lets be realistic here and maybe come back to this thread in 8 years time, when Rodgers has a similar number of years & games under his belt. Lets see then what his stats and ratios are.

In terms of career achievements, obviously not - Manning has 16.5 years under his belt (having missed one) and Brady has 12.5 (also having missed a full year) whereas Rodgers has 6 (having missed half of last year).

What I was referring to was what their levels of play in their careers. It is madness to say Rodgers is not comparable there, because since 2011 you simply cannot say any QB has ever played better. You can make arguments that QBs have played as well as well as he has, but not better.

However if you want to only compare on a bulk/year-for-year basis, Rodgers individual performance over his first six years starting has been on another level to what Manning/Brady were in their first six.

Manning, Brady, they've all led pass dominated offenses. They never had dominant unstoppable running games. So what if Rodgers hasn't a great running game. He's not the only one that's had that during his career. The Broncos currently have the 27th ranked RG in the league, The Packers 24th and the Pats 19th. And last year the Packers had the 7th best running game in the league, higher than the Broncos & the Pats.

The Packers are 25th and the Patriots 17th, but let's look back on other years...

From 2010-12 Aaron Rodgers was 60 yards off being the Packers leading rusher. In 2010 taking away the scrambling yards of all three running backs, the Packers had 1,250 yards at 3.5 per carry... the Colts had 1,464 at 3.9 per carry... the Patriots had 1,912 on 410 carries. In 2011 Manning missed the season, while the Packers had 1,301 yards on 3.9 per carry while the Patriots had 1,655 at 4.2 yards per carry, discounting scrambling. In 2012 the Packers had 1,443 rushing yards at 3.8 per carry while the Broncos had 1,826 at 4.0 yards per carry and the Patriots had 2,152 yards at 4.3 yards per carry again discounting scrambles. In all three seasons Rodgers' running backs finished below the Colts/Broncos and Patriots both in total yards and in yards per carry.

This season the Packers have 658 yards at 3.8 yards per carry (Lacy is a very good RB but has been misused a lot - he is poor out of shotgun, much better out of I-form/singleback type sets), the Broncos have 758 yards at 3.9 per carry, and the Patriots have 895 yards on 3.8 per carry. So the per carry percentages have been very close, but again the Packers have had less yards on the ground once you take QBs out of the equation.

Last season Rodgers finally did have a quality running game, but he missed half the season.

And when it comes to the O line? Rodgers almost doesn't even need it, he's always been a very good mobile QB. He can just run around, buying time and someone usually gets open. When a QB can improvise like that, the O line certainly won't affect him to the same extent as it would an immobile QB or classic pocket QB.

And he has over 1,100 rushing yards in the last four years. These are two vital factors as to why it really can't be said that any QB has ever performed on a higher level than Rodgers has since 2011. And the one year he had a solid offensive line in 2011, he had one of, if not the, best regular season in NFL history.

Billy86

Vandango wrote: »

1) Where was I talking about a running game?

2) Who said rushing isn't part of an offense?

Every team has a running game. And how did the running game, help Brady overcome losing his top 6 receivers and 94% of passes last year? And why then did the Pats run the ball less last year, than they did in 2012? In 2012 the Pats had more rushing yards and had 53 more rushing attempts than all of last year.

You said it - "Gets to the AFC title game last year with no offense or O line". When your rushing offence is top 10 in total yards and yards per carry, you clearly have something of an offence around you. It also helped Brady that when the WR/TE issue was at it's worst early on, the defence was for the most playing fantastic, allowing a very low 16.6 points on average with 14 turnovers in the first six weeks of the year - only eight teams managed more than than for the entire season. They fell off dramatically, but at least some guys like Gronk were back by then and others had adjusted better to the system.

They ran the ball less for two reasons - first, the had less total plays (1,164 in 2012 vs. 1,098 in 2013) and second, because they were behind or not comfortably ahead far more often - in 2012 their point differential was +226 whereas in 2013 it was +106. In 2012 the Patriots passed 55.06% of the time compared to 2013 when they passed 57.19% of the time.

poldebruin

...I know it's mentioned a lot about the recent passing rules changes. But just in case you were wondering just how much of an effect it has.
The difference between playing circa 1998 and 2014

Obviously as you go back the rules against passing are tougher, no protection for QBs, late hits, blows to the helmet, going low on QBs were all pretty much allowed. And that's not taking into account how receivers were treated. That is why Dan Marino's 1984 season stands out to me as being the greatest single season, statistically speaking.

Billy86

poldebruin wrote: »

...I know it's mentioned a lot about the recent passing rules changes. But just in case you were wondering just how much of an effect it has.
The difference between playing circa 1998 and 2014

Obviously as you go back the rules against passing are tougher, no protection for QBs, late hits, blows to the helmet, going low on QBs were all pretty much allowed. And that's not taking into account how receivers were treated. That is why Dan Marino's 1984 season stands out to me as being the greatest single season, statistically speaking.

It's definitely a strong possibility alright - Steve Young in 1994 (which also incldued 293 rushing yards and 7 TDs) also deserves mention, as does Joe Montana in 1989 and Otto Graham in 1953 (though the game was unrecognisable then - Don Hutson in 1943 also deserves mention for WRs, that's an all pro season even 71 years later).

Of course on the flipside, I think it gives a good argument for all of Woodson, Bailey, Ed Reed and probably Polamalu being among the very, very, very best DBs of all time on the flipside. Not saying you are at all, but I can't stand seeing the same people who talk about how QBs were so much better years ago because of the rule differences write off these guys in favour of Ronnie Lott, Night Train Lane, Deion Sanders, etc by default... which happens all too often really.

A complete hypothetical of course, but one of the great "whatifs" in my opinion would be the same team having drafted all four of those guys - they came within six year of each other, and though Woodson (4) and Bailey (7) went high overall, Polamalu (16) and Reed (24) were very obtainable... and the four of them really couldn't possibly compliment each other any better if they tried, in terms of style of play and strengths of their games.

raze_them_all_

eagle eye wrote: »

Rodgers is a great QB but he has a long career ahead of him unless he gets a bad injury. When he starts to slow down a bit and becomes less elusive we will see exactly where he stands in the GOAT conversation. Right now he is a couple of superbowls away from earning the title.

same amount as manning though :pac:

poldebruin

Actual vs Expected Win Expectancy for QBs



Some interesting things jump out on these charts.

You can really see who the gunslinger type QB's are, Luck, Romo and Stafford
Stafford wins less than 50% of games when he has a 75% win expectancy, which is obviously poor, but wins over 25% of games when he has less than 10% Win Exp. It would make me wonder if he suffers from the type of play calling that is too conservative and holds him back until the game is out of hand.

Luck's chart is similar in games he has a low Win Exp. Over 50% wins in games he has less than 10% Win Exp. which is insane and he doesn't suffer the same dip that Stafford does when the games are in the balance. This stacks up well with what we know already about Luck and we have seen him pull some great comebacks in his first 2.5 years.

Romo shows the value he adds to the Cowboys, with some great unexpected wins, but is one of the few QB's that actually come up short in games with 100% win expectancy. This also fits with the narrative we are frequently fed about Romo.

Brady, Luck and Manning stand apart, in that they add value no matter what the situation. Great it low Win Exp situations and lights out when the games are in the balance, while also being good closers. For example when Brady or Manning have a 25% win exp they win 50% of the games. When they reach 50% win exp they win 75% of the games.

Full article here in Skeptical Football Column

Love this line:
To start with, Peyton Manning not only wins a larger percentage of games than he is “supposed” to, but he exceeds that rate with just about the most perfect curve of anyone in the entire data set. It’s as if his game is starting to approach a mathematical ideal of good quarterbacking

Poor Ryan Fitzpatrick starts as expected and wins when he;s supposed to (close to 100% as you would expect of all the QBs) but anything in between and he's below average.