Writers of Pro Football Prospectus 2008

Most Recent FO Features

MasonHut15.jpg

» FEI Week 3: Bulldogs Bog Down

Georgia had a significant field-position and turnover-value advantage over South Carolina, but couldn't convert with the game on the line.

16 Mar 2007

Opposition Matters

by Bill Barnwell

These are heady times at Football Outsiders. During this year's playoffs, we introduced a free preview of the premium stat database that will be out in full force sometime late this summer. One of the reasons that the split database is pretty exciting is that it will have 11 years of data, increasing our data points for DVOA and basis for analyses by more than 20 percent compared to a year ago. We're welcoming friends of varying levels of depravity: the Houston Oilers, Sean Salisbury, Bernie Kosar, Willie "Big Play" Clay... All fun, new additions to the FO family.

(Yes, 1996 is done. We'll have the numbers and commentary up on the site sometime this summer, once the book is finished. For now, I'll just say that Green Bay dominates the league and these numbers will finally kill the myth that DVOA doesn't like Barry Sanders. -- Aaron)

The other advantage with having 11 years of data to analyze is that splitting that data up and then looking at the cumulative results provides more meaningful numbers. With quarterback numbers, for example, two more seasons' worth of games provides some muscle to some otherwise-flimsy splits.

In the new DVOA era (1996-2006), there have been 5,807 instances of a quarterback throwing 10 passes in a game -- this definition is used as opposed to starts because there are times where a starter throws three passes and gets knocked out of a game, or throws eight passes because he's Kyle Orton and can't be trusted with a remote control and a bottle opener, let alone an offense. There have been 5,536 starts by quarterbacks over that time frame, which would leave around 271 times, give or take a few, where a quarterback threw ten passes and left the game in the place of a backup who did the same thing.

These quarterbacks have thrown for nearly 1.2 million yards. They've posted ten perfect games, according to the official NFL passer rating: Peyton Manning thrice, Kurt Warner twice, Ben Roethlisberger, Chad Pennington, Trent Green, Doug Flutie, and Kerry Collins. There have also been 13 devoid of any value whatsoever: stand up Kent Graham, Scott Mitchell, Trent Dilfer, Anthony Wright, Dave Brown, Jeff Garcia, Eli Manning, Tony Graziani, Joey Harrington, Randy Fasani, Ryan Leaf, Rex Grossman (Week 17 against the Packers), and Tim Hasselbeck. Let's look at them judged against two variables, to see how a quarterback's opposition affects his play.

Opposition Wins

How do quarterbacks perform when they play against excellent teams? On one hand, it would follow that they throw more because their team is more likely to be behind; on another, that their team would want to run the ball to keep the other team's offense off of the field. If you're Goro, you can take on two more hypotheses, neither of which would be particularly supported by FO: Since the opposing team is obviously a winner, they must be causing turnovers, which would seemingly indicate more interceptions; on the other hand, quarterbacks would be gunning up to play a team they know to be good, and you've gotta knock down the champ, more bombs get thrown, leading to higher yardage totals and more touchdowns. Only the best ex-Redskins quarterbacks can get away with using all of these in the same game, let alone the same drive, as he does. Oh well; one person's deft is another's daft, right?

Let's take a look at the numbers and see if any of those are true.

  GP Att Comp Comp % Yards Yd/Att TD INT QB Rate
<4 wins 810 27.4 15.8 57.72% 186.9 6.81 1.4 0.9 82.1
4.5 to 6.5 1148 27.8 16.0 57.58% 183.6 6.60 1.2 1.0 75.7
7 to 9 1945 28.0 15.9 56.93% 181.8 6.50 1.1 1.1 72.2
9.5 to 11.5 1051 28.7 15.9 55.47% 178.7 6.23 1.0 1.2 69.0
>12 wins 853 28.0 15.7 56.14% 175.6 6.28 1.1 1.3 67.6
TOTAL 1161.4 28.0 15.9 56.77% 181.3 6.48 1.2 1.1 73.3

The first column (games played) is the count of all performances playing against teams with that range of wins, while the other ones are all averages.

What the data shows is a steady degradation in every category. Completion percentage, yards, yards per attempt, touchdowns, and QB rating all go down, while interceptions go up. Is it enough to make a Voros McCracken-style hypothesis that a quarterback has little to no effect on his passing performance? Definitely not. The correlation coefficient between quarterback rating and the opposing team's wins on the season is -.18, not particularly strong. It does provide some evidence, though, that a quarterback's performance in a given game may have a lot more to do with the defense he's playing than anything about him in particular.

Opposition Pass Defense DVOA

Of course, team wins also aren't a perfect measure of how good a team is at affecting the play of the opposition quarterback. To focus on that, let's look at how quarterbacks fare against different sets of pass defenses.

  GP Att Comp Comp % Yards Yd/Att TD INT QB Rate
13.4%+ 1071 29.5 17.5 59.24% 205.2 6.95 1.4 0.9 83.3
4.5 to 13.4% 1304 27.8 16.1 58.00% 186.0 6.70 1.2 0.9 78.5
4.4% to -4.4% 1177 28.6 16.3 56.96% 183.2 6.41 1.2 1.1 74.4
-4.5% to -13.3% 1045 27.7 15.4 55.46% 174.4 6.30 1.1 1.2 70.4
-13.4%- 1211 27.2 14.7 54.00% 163.1 6.00 0.9 1.4 61.9
TOTAL 1161.4 28.2 16.0 56.73% 182.4 6.47 1.2 1.1 73.7

Note that some of the averages might be different due to rounding.

Interesting -- the same trend that we saw in the wins grouping rears its head again, but even more virulently. Against an excellent defense, as opposed to one below replacement-level, quarterbacks lose a full yard per attempt, half a touchdown turns into an interception, 40 yards disappear, and five percent fewer passes get completed. Those are some pretty dramatic effects. QB Rating shows a .28 correlation with pass defense DVOA, which is actually a negative correlation (since a negative DVOA represents a better pass defense, while a high QB rating represents a better passing offense), but not as strong as it might seem looking from that chart.

Let's think about fantasy football for a minute, too. Should you be picking your quarterback solely based upon the defense he's playing? On average, playing a quarterback against a replacement-level pass defense versus an upper-echelon pass defense would earn you five fantasy points. Of course, you can use the Football Outsiders' breakdowns of performance against receiver types and common sense to figure out whether a certain matchup is better or worse than the average, but five free points are nice. The correlation between fantasy points and defensive DVOA is -.25, so it's not strong enough to be a quick-and-dirty rule, but it's something to keep in mind those weeks that you feel uncomfortable starting Aaron Rodgers in Denver next season.

The research raises a pretty interesting, but difficult question: how much credit does a quarterback deserve for a good performance? Alternately, what about the defense? How much credit should they get? Is the question unanswerable?

Fun With Splits

You'll see some of these splits show up in the Pro Football Prospectus 2007 player comments, but here are a few tidbits to tide you over:

  • Brett Favre, he of the maximum 176 starts, is a great example of how performance changes by opponent. In his 23 games against teams with a replacement-level or worse pass defense, his QB rating is 98.8; he has a 3:1 TD to Interception ratio and completes 64% of his passes. Against the top pass defenses? 35 games, 71.8 QB rating, a 1:1 TD to Interception ratio, and he's completing 56% of his passes. He loses nearly 27 points in QB rating in the process.
  • The regulars who lose the most under the same switch? Drew Bledsoe and Peyton Manning, who lose 35.61 and 35.09 points of quarterback rating. It's amazing that some Boston writer didn't do this research first to point out that Peyton can't win in the playoffs because he doesn't show up against the great teams and beats up on the crappy ones, A-Rod style. Oh well. Too late now! Bledsoe has a particularly ugly 58.6 QB rating in 34 games.
  • Poor Shane Matthews. He got 25 games in this eleven-year period and 11 of them were against pass defenses who were better than -13.3%. Five of the aforementioned Graziani's nine career performances came against the same type of defense; there's a reason he's in the Arena League now.
  • Finally, want a consistent quarterback you can rely on? The guys with the lowest variances across the different splits were Trent Green, Jon Kitna, Gus Frerotte, and Jake Delhomme. No fun, you say? The guys with the wildest split differentials were Bledsoe, Drew Brees, Dan Marino, Ben Roethlisberger, and -- sigh -- Eli Manning. Give me a guy banging his head into the wall any day.

Posted by: Bill Barnwell on 16 Mar 2007

50 comments, Last at 08 May 2007, 5:24pm by Jack Neefus

Comments

1
by Peter (not verified) :: Fri, 03/16/2007 - 12:14pm

Mortal Kombat reference makes this the best article in years.

2
by McGaytrain (not verified) :: Fri, 03/16/2007 - 12:18pm

The 4 low-variance QBs strike me as guys who got lots of starts late in their career. Yay NFL farm system.

3
by dryheat (not verified) :: Fri, 03/16/2007 - 12:23pm

You might be mixing up the cause and the effect.

4
by Charles the Philly Homer (not verified) :: Fri, 03/16/2007 - 12:34pm

Despite the sleet, a sports media that despises its own team, and the joy of college basketball gambling, only this article has truly brought that midseason warmth back, if only for a few moments. Thanks.

5
by Rich Conley (not verified) :: Fri, 03/16/2007 - 2:02pm

"The guys with the wildest split differentials were Bledsoe, Drew Brees, Dan Marino, Ben Roethlisberger, and — sigh — Eli Manning."

That may not actually be a bad thing for Eli. Thats some pretty good company.

"Drew Bledsoe and Peyton Manning, who lose 35.61 and 35.09 points of quarterback rating"

Right, so basically, good and great quarterbacks perform well against good defences. Great QBs perform even better against bad defences. These stats are really making me start to think that if Bledsoe had been born 10 years earlier, he'd hold pretty much every passing record out there.

6
by Ben (not verified) :: Fri, 03/16/2007 - 2:06pm

I agree with 3. I wasn't suprised that QB performance declined as defensive DVOA improved. If you have a good pass defense, the opposing QB isn't going to have good days against you. In fact, DVOA already adjusts for this, the strength of the opponent.

I don't think splits tell a good story either. I want to know which QB has the best rating/VOA/PAR against the top defenses, and how much they are actually beating up on bad defenses

7
by miguelacero (not verified) :: Fri, 03/16/2007 - 2:09pm

In the intro, I think you mean Randy Fasani, not Fasano, the Stanford stiff who got a couple starts for Carolina, then was traded to the NY Jets in the offseason but opted to retire rather than play another worthless game.

8
by Pat (not verified) :: Fri, 03/16/2007 - 2:20pm

In fact, DVOA already adjusts for this, the strength of the opponent.

Yes, but quarterback rating doesn't, which is the entire point - you need an opponent-adjusted stat, because otherwise you're just staring at opponent effects all the time. Hello, Rex Grossman!

You'd be amazed how many football fans don't understand that teams play worse against good opponents. There's a reason the teams in the Super Bowl frequently look like crap.

So you get dumb comments like "look how much worse we did in the playoffs! good teams step it up!" No, everyone plays worse versus good teams, on average.

9
by dryheat (not verified) :: Fri, 03/16/2007 - 2:39pm

To clarify, in Post 3 I was referring to Post 2, which seems to imply that quarterbacks who get a lot of starts produce a low variance. I think it's possible that these quarterbacks get a lot of starts because of their low variance. But I really don't know.

10
by Sergio (not verified) :: Fri, 03/16/2007 - 2:43pm

Is the DVOA used for defenses season-wide, or game-specific? Yes, I like hypened-formed terms, thank you very much.

So Marino is very variable, at least 96-99... hmmm... fits the memories. Thank you once again, JJ...

11
by Blah (not verified) :: Fri, 03/16/2007 - 2:51pm

so quarterbacks play worse against better defenses! wow ...

12
by Pat (not verified) :: Fri, 03/16/2007 - 2:52pm

#10: The point is that the less-variable QBs seem to be the ones who didn't get starts early in their career. Green, Delhomme, Kitna - all didn't become fulltime starters until their 3rd season or later. Plus, Kitna was 27, Green 28, Delhomme 28. (Frerotte was younger).

13
by Mike Leach (not verified) :: Fri, 03/16/2007 - 2:53pm

Who the hell is Barry Sanders? First!

14
by Mike Leach (not verified) :: Fri, 03/16/2007 - 2:54pm

Or not!

15
by johnt (not verified) :: Fri, 03/16/2007 - 2:56pm

With Favre and Bledsoe I'm pretty sure those "giant splits" are the product of some truly awful games, especially since you gave the high/lows. With Manning I'm not so sure since you didn't give the actual range. Are we talking a rating of 60 to 95 Bledsoe style or some 80-115 type situation? If the latter the reason the column's never been written would be that it would be a pretty dumb thing to write a column about how "Manning is superhuman against bad defenses... but only above average against the best defenses in the NFL!". Not that that would stop Ron Borges, but......

16
by Reinhard (not verified) :: Fri, 03/16/2007 - 2:57pm

I think that the research here is incomplete. Quarterbacks do worse against good pass defenses... yes, by definition! Otherwise they wouldn't be good pass defenses.
The more interesting area is to measure HOW MUCH worse or better than expected that they do.

17
by Larry (not verified) :: Fri, 03/16/2007 - 3:48pm

Yes, the vs. Defenses by PassDef DVOA has to yield exactly that result because you've just added up the plays that determine the Passing DVOA rating in the first place. The observation of the existence of trends is trivial in the extreme. All it tells you is the conversion formula between DVOA and Passer Rating.

Also, with the number of games played near 1000, values should have at least a second decimal place.

Better would be to take something like 'The Best QBs v. The Best Overall Defenses' and the reverse. Then at least the data you are looking at isn't the same data that sets up the categories. Even the first table suffers from this, since the less than 4 wins category is mostly data from games lost. Games lost look worse than the data from Games won in the greater than 12 wins category. Unless you believe passing yards don't correlate with winning at all, which would be intersting if true.

To summarize, this analysis has cause and effect completely reversed. The first table shows how pass defense correlates with winning. The second shows how DVOA is calculated. Niether is especially surprising, though perhaps that's the point.

18
by Pat (not verified) :: Fri, 03/16/2007 - 4:08pm

Yes, the vs. Defenses by PassDef DVOA has to yield exactly that result because you’ve just added up the plays that determine the Passing DVOA rating in the first place

Only if the pass defense DVOA rating is only for that game, which I doubt it is. I'd imagine it's for the full season.

Also, with the number of games played near 1000, values should have at least a second decimal place.

Not if you really want to be correct. Four significant figures is too much (that'd be a part in 10,000). Should be three (That means TD/INT should go out more, and yards/game and completion percentage are too precise).

The first table shows how pass defense correlates with winning.

Only if "passer rating" is correlated with "pass defense", which is the entire point of the article - that quarterback rating is dependent upon the opponent, not just the quarterback.

19
by Will Allen (not verified) :: Fri, 03/16/2007 - 4:42pm

This is good work, even if it is not surprising in the least. Anybody who pays attention to athetic competition with an analytical approach becomes aware that the superior athletes really tend to pad their stats against bad competition, and do less well against good competition. Yeah, there are some outliers. Michael Jordan went off for more than 60 in the playoffs against a championship caliber Celtics team, for instance. For the most part, however, guys get to the various Halls of Fame by annihilating poor competition. Think of it as an aspect of GUTS vs. STOMPS.

To bring up another non-football example, I attended game 6 of the 1991 World Series, a classic nail-biter in a Series of nail-biters. In the 11th, Puckett was the first batter, and when the P.A. announcer introduced Charlie Liebrandt as the Braves' pitcher, I turned to my friend and said, "This game is over". It was, because Liebrandt at that time in his career was getting by purely on shoveling up left-handed crap, hoping to befuddle hitters, and while Puckett was a good hitter against anybody, he was an absolute total destroyer of left handed softly thrown crap. Twenty seconds later, Puckett was circling the bases.

Great athletes make their fortunes by annhilating much poorer athletes, not by "stepping up", or being "clutch" against similarly skilled competitors.

20
by Bob (not verified) :: Fri, 03/16/2007 - 4:52pm

Seems like an advertisement for the FO Premium Stats database. I wonder how much this is going to cost.

21
by SJM (not verified) :: Fri, 03/16/2007 - 5:05pm

While I disagree with most of post 18, I think he's right that a "Best QB's vs best defenses" article with variations like best QB's overall, best rise-to-the-challege QB's, and so on with the defenses, would make an excellent follow-up.

22
by Chris (not verified) :: Fri, 03/16/2007 - 5:10pm

9 Pat- No joke. I'm not really one to listen to the pre-game shows but it seems like every time I turn one on it's " Will we see good Rex or Bad Rex"?

The guy takes chances ( risk), which will lead to a higher Variance. It's not that complicated.

Fantasy football is all about match ups at the RB position. Remember the whole " start the RB of the team who is playing the Colts each week, no matter who he is" argument?

Will Allen- Speaking of picking up cheap stats, wasn't that a lot of what S. Alexander did 2 years ago in seattle? The Hawks had an easy schedule and Alexander picked up a lot of garbage yards against the Arizonas and the other easy teams they played. But I remember he had maybe 75 yards against the Jags and he didn't do as well against the other "respectable" defenses.

So if Peyton Manning is throwing for 300+ yards and 3-4 TDs against Houston, but say only 229 against the Bears...

I think Eli's lower rating against good defenses could be because we don't have THAT much data on him and a lot of his very first ( and worst) starts were against good teams, and his most recent starts were when his whole team was injured. The Giants schedule wasn't easy last year and mix that with the injuries and he's playing uphill.

23
by Will Allen (not verified) :: Fri, 03/16/2007 - 5:33pm

Chris, the only stats I consider "cheap" are those obtained against a defense which really isn't trying to defend against what the offense is doing. Running the ball in the fourth quarter while behind 4 touchdowns, for instance. When a guy is piling up stats against a bad team which is doing it's best to stop the guy, I don't think it should be seen as cheap. A semantic difference, perhaps, but one that should be noted.

24
by Vince (not verified) :: Fri, 03/16/2007 - 5:51pm

First things first: This is the kind of stuff I love to see at FO.com. I can get interviews or talent profiles anywhere; I come here for statistical analysis.

This seems like a good introduction to a study rather than a study itself. You've shown that all QBs put up better numbers against bad teams/defenses then against good teams/defenses. I guess maybe that's not self-evident, but it seems that way to me.

The next step seems to me to see how good/bad/mediocre QBs perform against good/bad/mediocre teams/defenses.

Finally, I think the real question you're getting at here is "Which is MORE important: The QB or the team he's playing against?" Obviously, the results on the field will reflect the quality of the QB and the defense, but is that a 50-50 split? Probably not; my guess is that it would be around a 60-40 split, with the QB mattering more than the defense, but that's just a guess.

I would suggest running a regression formula with the known Xs being QB DVOA and pass defense DVOA, and the known Ys being either DPAR or DPAR/attempt for each game. That should answer the question somewhat.

25
by Chris (not verified) :: Fri, 03/16/2007 - 6:40pm

Yeah, but haven't you ever seen a stadium empty and the players just not care as much?

A 1st and 10 with 6 minutes to go in a hopeless game is different than 3rd and 1 of a key possession.

I'm more impressed with a guy that can grind out 100 against anybody, and played better competition than a guy who rushed for say 180 against the Cards and other weaker teams, and 75 against the Jags and better teams.

26
by Jerry (not verified) :: Fri, 03/16/2007 - 7:18pm

This would be more interesting as a 5-by-5 matrix; does a second-quintile QB do better against a second-quintile defense than a fourth-quintile QB does against a fourth-quintile defense?

27
by kibbles (not verified) :: Fri, 03/16/2007 - 7:24pm

[broken record]

So... any chance of getting the single-season DVOAs by position prior to 2000?

[/broken record]

Seriously, though... I know that this is a free service (and quite an amazingly awesome one at that) that you guys are providing, and I have no right to complain. It wouldn't bother me if you didn't have the data, but the knowledge that you've had the data compiled and run and just sitting there for years without any way for us to access it just drives me a little batty. I don't want you to think of me as a discontented fan, just as an avid fan who can't get enough of your stats. It's like statistical heroin, and now that I'm addicted I need my fix.

28
by Pat (not verified) :: Fri, 03/16/2007 - 7:51pm

This would be more interesting as a 5-by-5 matrix; does a second-quintile QB do better against a second-quintile defense than a fourth-quintile QB does against a fourth-quintile defense?

Without even looking at the numbers, I'd say "almost definitely."

Quarterback is one player. Defenses are 11. Naturally, the total variation in quarterback play is going to be larger than the total variation in a defense's play, simply because it would take serious effort to put together 11 players that were as bad as, say, Ryan Leaf.

29
by J.D. (not verified) :: Fri, 03/16/2007 - 10:20pm

I think I shall root against the Eagles this season.

30
by Larry (not verified) :: Sat, 03/17/2007 - 1:15am

Only if the pass defense DVOA rating is only for that game, which I doubt it is. I’d imagine it’s for the full season.

Almost every play against that defense is included in the numbers in the table [QBs throwing less than 10 passes are excluded and the denominator is slightly different for games where 2 QB thre 10 passes, but these effects are almost certainly very, very minor - about 5% according to the article]. So, it is in fact the entire set of numbers that define the DVOA in the first place.

[Table 1 shows pass defense correlates with winning o]nly if “passer rating� is correlated with “pass defense�, which is the entire point of the article - that quarterback rating is dependent upon the opponent, not just the quarterback.

The stat lines shown in table 1 are the ENTIRE stat line against teams with those records (except for QBs throwing less than 10 passes in a game, but that's a pretty minor difference). So, you have the passing numbers given up by "Teams with less than 4 wins", "Teams with 5/6 wins", etc. For example, the second line is teams that went 5-11 or 6-10 (or +/- half a game for a tie). There are 1148 games in that row. Those games represent 71 teams (plus a few games where 2 QBs played) and their record was 311-745, more or less, in those games. So, I'm just looking at the statline of a bunch of games where I know the record of those offensive passing statistics is 745-311 (more or less) and the record of the defenses that gave up those numbers was 311-745 (more or less). The next line the records move towards .500. And so on. Have we really learned anything here other than that the passing stats of teams that win are better than the passing stats of teams that lose? I would say not from this analysis.

31
by D (not verified) :: Sat, 03/17/2007 - 1:29am

This article itself was alright but the statement about the '96 Packers is what really piqued my interest. I know that they are the only team since the '72 Dolphins to lead the league in both most points scored and fewest points allowed and leads me to believe we may have a team that can take the title of "Greatest Team in DVOA History" from the '99 Rams.

32
by Steve Greenwell (not verified) :: Sat, 03/17/2007 - 6:13am

If you’re Goro, you can take on two more hypotheses...

But what if you're Shang Tsung? Does he only get two hypotheses, or does each form get its own hypotheses?

33
by Kerwin Nagy (not verified) :: Sat, 03/17/2007 - 11:09am

re: 34

I think the 1989 49ers need to be looked at, too.

34
by Jerry (not verified) :: Sat, 03/17/2007 - 1:46pm

Re #29:

Without even looking at the numbers, I’d say “almost definitely.�

And without looking at the numbers, I'd say that passing against with teams with good records or good pass defenses is harder than passing against teams with bad records/defenses.

35
by D (not verified) :: Sat, 03/17/2007 - 2:42pm

#36
I was just refering to teams from the DVOA era. The play-by-play isn't readily available for years prior to 1996. But yes, the '89 Niners were a great team and might have a higher DVOA than the '99 Rams.

36
by stan (not verified) :: Sun, 03/18/2007 - 1:30pm

The assumption which underlies this analysis is stupid.

If my QB drops back to pass, has time to throw and an open receiver, he should complete the pass. If he gets no time to throw, I hope he is good enough to throw it away rather than get sacked (even if it hurts his passer rating). If he has no one open, same thing.

What does the lower passer rating against the better defense tell me about the quality of his play? Absolutely nothing.

If his offensive line gives him time (even against the best pass rushes) and his receivers get open (even against the best coverages), what does it tell me about his passing ability, if he gets the same rating against the bad defense as the good?

Answer -- absolutely nothing. If he has time to throw and open receivers, he SHOULD have the same numbers.

37
by admin :: Sun, 03/18/2007 - 7:54pm

No Brady-Manning arguments allowed. Please take your Kevin Curtis talk to the NFC East thread on the new discussion board. Thank you.

38
by SJM (not verified) :: Sun, 03/18/2007 - 9:41pm

Since my first post ripping another poster for illogic got deleted, let me respond to stan:

The underlying assumption here is that there are a lot of contradictory bits of conventional wisdom, and some of them are wrong.

Passer rating is not a great measure of QB play, but it's OK. PR does not penalize incompletions all that much, compared to the influence of TD passes and interceptions. If a QB has a lower passer rating against good defenses, it PROBABLY means he didn't perform as well as he usually does, regardless of any penalty he incurs for taking an incompletion instead of a sack.

Of course an offensive line can beat a great pass rush, and receivers can get open against a great secondary, but a QB can still be affected by the defense even if those things happen (if they are good at disguising coverages and blitzes, they can prevent him from getting comfortable or finding his open guys).

So what is the point? The point is that all QB's, even the great ones, pad their stats against patsies and aren't so great against great defenses (they don't "get up for the big game.") Even though it seems obvious, I think that is a valid lesson, and one which the general public probably hasn't learned. Another lesson is that good passing defenses depress passing numbers even though the opponent is often behind and wants to pass more.

In any case, in an earlier post I mentioned that I want to see a follow-up which compares the best QB's to each other when facing top D's, to see if some actually do rise up (or at least play up to standard) and others wilt. I think that could be very interesting.

39
by throughthelookingglass (not verified) :: Sun, 03/18/2007 - 10:12pm

It's the return of unintentional responses to future posts!

40
by Dean from Oz (not verified) :: Sun, 03/18/2007 - 11:09pm

Re13:

What is the latest serious call for "First!"?

11 posts in, 2 and a half hours after the actual first...thats not bad.

41
by Lou (not verified) :: Mon, 03/19/2007 - 5:01am

Is the opposite true? How much do good defenses pad their stats against bad offenses?

42
by PackerNation (not verified) :: Mon, 03/19/2007 - 8:50am

This research supports my own findings that Brett Favre doesn't play particularly well against quality defenses and hasn't for a while. Of course, no quarterback plays as well against good defenses.

I'd be interested in seeing some type of ranking of how various QBs have done against top echelon defenses over the years.

The findings that Manning doesn't play particularly well against better defenses explains a lot regarding the lack of success he'd had in the playoffs up until this year, IMO

43
by Wanker79 (not verified) :: Mon, 03/19/2007 - 9:28am

This is one of those articles that I'm going to file away under "Support for Arguments with Idiots". Every single person who has complained that this is just all common sense has obviously become spoiled by the high level of intelligent analysis by the writers and posters here at FO.

Of course this seems fairly obvious to most of us. Sadly though, we are still the minority. This article goes right next to the ammo for my inevitable "fumble recovery is mostly luck" and "tackle totals are meaningless" arguments.

44
by countertorque (not verified) :: Mon, 03/19/2007 - 7:11pm

Nice article.

I was expecting QB attempts and yards to go up vs. good defenses, but they both go down. I thought that teams that are behind pass a lot more to try to catch up, but these tables seem to say that they don't.

45
by Wanker79 (not verified) :: Tue, 03/20/2007 - 9:17am

Re: 44

My guess would be that the percentage of plays a team passes may go up, but the better defense forces shorter drives thereby limiting the total number of attempts. But like I said, that's just a guess.

46
by SJM (not verified) :: Tue, 03/20/2007 - 8:44pm

Wanker,

That makes a lot of sense and is probably the correct explanation. In addition, a team that passes all the time will face a lot more 3rd-and-long situations than usual, which will also shorten their drives.

47
by Strange/David (not verified) :: Wed, 03/21/2007 - 12:37pm

Re: 37

Around here, the only Brady-Manning argument we allow is "Which would be a better name for a character in a book: Tom Peyton, or Brady Manning?"

For Colts fans (like me), the only sad part is that we can't find a way to put Peyton first.

(Incidentally, I'm voting for Brady Manning.)

48
by stan (not verified) :: Fri, 03/23/2007 - 3:38pm

If you want to do something interesting, do a stats analysis on two topics:

1) Defenses that feast on bad teams and get ripped by good offenses. This often happens with certain college teams which use a lot of press coverage and blitz like crazy. Against a team that can't pass protect, they can put up a shutout and limit yardage to ridiculously low levels. They overwhelm the run and hound a QB to an early shower. Of course, the DBs don't have to cover very long. This type defense that feasts on patsies will have awesome stat rankings. At least until they run into a team that can protect, has receivers who can get open and a QB who can throw well. Without the suffocating pass rush, the DBs start looking horrible and the defense can end up torched. In essence, they are using a feast or famine, roll the dice type strategy. (Example, Va Tech a few years ago which completely overwhelmed offenses all season until running into a good Miami passing attack and got hammered.)

Compare and contrast to a team that plays vanilla. Often with a good offense on their side, they know that they simply need to avoid the big play to win the game. They will give up the short pass, make the opponent go the long way and will yield more field goals and fewer TDs. Even bad offenses will be able to move the ball. Stats won't look as good against bad offenses. But will often yield similar numbers against good offenses.

My point -- might we use the stats as an indicator to point out possible differing playcalling strategies and make projections for future matchups? I.e. is someone's outsized success against the weaker part of the schedule a marker for trouble against better offenses.

2) Another area for analysis that is counter to the first assumption in this article. Isn't a QB who fails to torch bad defenses, yet plays about the same against good defenses, showing that HE is the problem that is keeping a good offense from being great?

Let's assume we have a QB who is generally regarded as pretty good. It's just that he has a bad habit of regularly missing wide open receivers, even when he has plenty of time to throw. Purely hypothetical, now. When he plays against bad teams, he has good numbers, but never seems to put up the huge passer ratings that other top QBs put up -- maybe because of this bad habit of missing open throws once in a while. Now when he plays against the better defenses, he still gets time to throw to open receivers because of the quality pass pro and legit run threat. So he still puts up good numbers, because his offensive teammates continue to give him the opportunity to do so.

The similarity of passer ratings (that is, the smaller divergence)when playing against vastly different quality defenses should be exactly what we expect of a mediocre QB surrounded by a quality offensive line and receivers who consistently get open. His ratings would not vary as much because his opportunities are so similar. His ratings don't go into the stratosphere against the bad defenses, because he lacks the consistency to take advantage of them.

49
by 18to88 (not verified) :: Sat, 03/24/2007 - 10:59am

Interesting stuff. Just look at Manning's numbers in 2006 vs. his 2006 playoff numbers. That says it all. Obviously wins are the name of the game and sometimes your worst statistical nights (the Ravens victory) are your finest hour. I know that sounds weird coming from a Manning fan. Don't worry we still love staring at his stats too.

Oh and Peyton's on SNL tonight.

50
by Jack Neefus (not verified) :: Tue, 05/08/2007 - 5:24pm

What I find surprising is how little the averages change for most quarterbacks. From best to worst categories by record, the completion percentage only goes from 56.1% to 57.7%, while the rating goes from 67.6 to 82.1.