One of the Sunday traditions on NFL Countdown was to list the 300 yard passers and 100 yards rushers from each week of action, and then point out how more frequently the 100 yard rushers won. It was a prime exhibit for those who get their correlations crossed with their causations to use in order to point out that establishing the run is what wins in football. If you've been paying attention this year, you may have noticed that even that old canard about 100 yard rushers versus 300 yard passers has disappeared. I'm going to show you some astonishing numbers. Here is the W/L record when a player has 100 yards rushing, or 300 yards passing, going back to 2002 (oh, and curse you Donovan McNabb and Tommy Maddox).

	300 yard passers				100 yard rushers
==========================================================================
	W	L	PCT		W	L	PCT
==========================================================================
2009	41	18	0.695		45	19	0.703
2008	47.5	28.5	0.625		100	30	0.769
2007	43	38	0.531		105	37	0.739
2006	39	29	0.574		112	47	0.704
2005	29	33	0.468		115	23	0.833
2004	36	45	0.444		135	44	0.754
2003	29	31	0.483		107	44	0.709
2002	44.5	34.5	0.563		96	40	0.706
==========================================================================

Those numbers include six shootouts in which both teams had a 300 yard passer, as well as two games where each team had a 100 yard rusher, and one game where one team had two rushers reach that mark. In games where one team has a 300 yard passer and the other doesn't, the record for the 300 yard passer is 35-12 (.744), compared to 42-17 (.712) in games where a team has at least one 100 yard rusher and the other does not.

Up until last year, teams that had a passer reach 300 yards won about half the time. The numbers for 100 yard rushers have held steadily north of 70%. Last year's mark of 0.625 was the highest since 1962 (NFL), when passers went 15-5-1 when throwing for 300 yards, and this year's percentage is exceeding last year. Last year, the 47 games won by teams with a 300 yard passer were the most ever. And so far this year, we are at 41, and every team still has seven games left. To put that in historical perspective, from 1970-1977, quarterbacks who threw for 300 yards won a total 45 games, over 8 seasons. We are almost there in nine weeks.

Weather should slow that pace down, but judging by the ratios from the last seven years (on average, there were forty-three 300 yard games from 2002-2008 through game 9), we are looking at somewhere in the range of ninety-eight 300 yard passing games in 2009. That would be the highest total ever. The previous high (90) was in 1995, the expansion year when Carolina and Jacksonville joined. That total dropped immediately to 50 the next year, so it was a freak anomaly partially aided by the expansion dilution year. The two other spikes in the percentage of 300 yard games were also directly attributable to either profound rule changes or massive expansion. The rule changes in 1978 to assist the passing game resulted in a four-fold increase in 300 yard games within a few years. The formation of the AFL and expansion from 12 to 22 professional teams led to a spike in high passing yardage games in the NFL from 1961 to 1963.

This one seems different. We are eight years removed from the last expansion. We are five years removed from some rules changes. I know Chase has some thoughts on this that he might write up. What do you think? Anomaly, or definite trend?

Yesterday, I argued that defenses shouldn't be thought of as good against the pass or good against the run; defensive statistics should be considered fluid measures, because most defenses can choose how they'll let the offense beat them.

So is this true? Is this even possible to measure? Assuming that defensive disparities (i.e., being very good at defending the run or pass and very bad at defending the pass or run) are random, then we shouldn't see teams consistently being good (or bad) in one just area of defense. One way to measure this is to break out each team into two half-seasons. To avoid injury issues, I spit teams into "odds" and "evens"; games 1, 3, 5, 7, 9, 11, 13 and 15 go into one group, and games 2, 4, 6, 8, 10, 12, 14 and 16 go into another. We would expect a metric like quarterback yards per pass to be relatively consistent in "odd" and "even" games, because averaging more (or fewer) Y/A than average is a repeatable skill; but we wouldn't expect teams that did well in even weeks on opponent missed field goals to again do well in odd weeks, because that's (generally) not a repeatable skill. So, are defensive disparities narrative statistics that merely described what happened or are they repeatable and predictable events?

I looked at every team-season since 1988, giving us 21 full seasons worth of data. I measured run defense by yards per carry allowed (relative to league average) and pass defense by adjusted net yards per attempt allowed (relative to league average). I also noted the raw number of points allowed by each defense (relative to league average). This gave me run, pass, and scoring defense grades for each half-team in the study. Assuming that defensive disparities were real and consistent, we should see them for teams in both the even and odd splits. If defensive disparities were merely a mirage, and defenses really force offenses to pick their poison, then there should be no correlation between the even and odd splits.
(Continued)

I've got a theory that there aren't necessarily teams that can only play the run well or can only defend the pass; rather there are just good defenses and bad defenses (and everything in between). Certainly there are teams that have played well against just the run and not the pass, and vice versa; there's no denying that looking backwards, certain teams in certain years were great in one area of defense and terrible in the other. But looking backwards, there are teams that recover lots of fumbles and teams that faced opposing kickers who missed a high number of field goal attempts, but that doesn't mean that those things will happen again in the future.

My claim is that theoretically -- i.e., if we played 1,000,000 games so that randomness would be eliminated -- defenses should just be defenses. No good run-D, bad pass-D teams; no average run-D, excellent pass-D teams; just defenses. That (I think) is a bold claim, so I better have some good reasons behind it. What are they?

1) Defenses are like chains: they're only as strong as their weakest links (for the flip side to this argument, see Brian Burke's article that offenses are like chains). Picture an unbelievable run-D teamed with an awful pass defense. That defense isn't going to be very good, as almost every team in the league could pass on them all day long. Flip the script, and nearly every team could control the game with power football against a defense that can't stop the run. On defense, if you have a weakness, almost every opponent can exploit it.
(Continued)

In the comments of yesterday's post about the best SRS teams of the 2000s, loyal PFR reader MattieShoes took issue with lumping together all incarnations of a team into one decade-long SRS calculation:

I don't think a decade-long SRS calculation is very good though -- Take the Raiders. They had some very good teams early on and some very bad teams more recently, so they average out to be a bad team for the decade. But now every team that lost to them in 2000-2002 is being adjusted as having lost to a BAD team, which isn't right.

I think it'd make far more sense to calculate yearly SRS and average those, since team strength is less likely to change drastically over the course of a season, and any errors caused by teams changing strength over the course of one season are limited to just that, one season.

I agree with that -- it probably won't impact the ratings much, but the method really does need to be changed to reflect the strength of each opponent when a team plays them, instead of using their decade-long strength rating. So here's the new deal: I calculated SRS for each season of the 2000s unto itself, and then took a weighted average of those SRS scores (by game played per season) to find the decade-long SRS rating. The results:

(Continued)

We've done a Players of the Oughts (or at least 90% of them) post, so I figured it was time to put out a preliminary list of the teams of the decade. The ranking method? SRS, of course. I used all games, regular-season and playoffs, and accounted for home-field advantage by adding 2.66 points (the average home-team margin of victory this decade) to the margin for away teams & subtracting the same amount from home teams.

And no, playoff games are not arbitrarily weighted more than regular-season ones, other than the fact that they will boost your strength of schedule score. I'm certainly open to different weightings, but they have to be grounded in something more legitimate than "I feel like playoff games should count five times as much", or whatever.

Anyway, here are the best NFL teams by SRS from 2000 through yesterday's games:

Yesterday, the Jets traded with the Browns to acquire Braylon Edwards. Edwards joins a relatively short list of notable wide receivers to switch teams in mid-season. Since the merger, there have been only 14 wide receivers to play for multiple teams in a season and to average at least 35 receiving yards per game for one team. Only 26 WRs have averaged at least 25 yards per game for one of the two teams. A complete list of those receivers is listed below:

year   team     name	         g      rec     yd      td      recyd/g

2008	det	Roy E. Williams	 5	17	232	1	46.4
2008	dal	Roy E. Williams	10	19	198	1	19.8

2007	mia	Chris Chambers	 6	31	415	0	69.2
2007	sdg	Chris Chambers	10	35	555	4	55.5

2006	rai	Doug Gabriel	 3	 5	 84	0	28.0
2006	nwe	Doug Gabriel	12	25	344	3	28.7

2005	car	Rod Gardner	10	 9	 84	1	 8.4
2005	gnb	Rod Gardner	 2	 4	 67	0	33.5

2004	dal	Antonio Bryant	 5	16	266	0	53.2
2004	cle	Antonio Bryant	10	42	546	4	54.6

2004	cle	Quincy Morgan	 6	 9	144	3	24.0
2004	dal	Quincy Morgan	 9	22	260	0	28.9

2004	rai	Jerry Rice	 6	 5	 67	0	11.2
2004	sea	Jerry Rice	11	25	362	3	32.9
 (Continued)

Reader Vishal Desai sent in a question this week wanting to know how uncommon it was for a team that rushed for 200 yards to lose, noting that both Dallas and Miami lost primetime games in week two. Well, that's what we are here for, especially when you compliment us.

Going back to 1978, and not including 2009, a team has rushed for 200 or more yards on 1,110 occasions. 135 of those teams lost the game. So, the short answer is, about 12.2% of the time.

A cursory and unscientific glance at that list of 135 losers provides some of the reasons why teams that rush for 200 yards lose games.

1) They really stink at passing the ball;
2) They are playing a team that is really good at passing;
3) The defense is a sieve;
4) They are a pretty good team that had an off day with turnovers;
5) They have a running back named Walter Payton(Okay, somewhat related to point #1, but more specific);
6) They are the only team in the modern era to run the Wing-T; or
7) They lose the game late, sometimes on a fluke play, like this one.

Other random notes:

--In the "thanks for opening up that old wound, Vishal" category, the Kansas City Chiefs lost 4 games from 2001-2004 in which they ran for over 200 yards, including this gem where they had over 550 total yards and still lost, and fooled me into thinking that Koren Robinson was going to be special.
--As further proof of the greatness of Payton, the Bears lost 7 games from 1978-1984 in which they ran for 200 yards. I also detailed this spectacular game for Payton in 1983 in a podcast, which was not quite a 200 yard game but was a loss.
--21 teams lost two games in the same season in which they had 200 or more rushing yards in both losses. Only the 1978 Kansas City Chiefs had more than two. They lost 5 games in which they topped 200 yards rushing. That was Marv Levy's first season and he installed the Wing-T Offense to try to protect a defense that had 8 starters age 24 or under, and rookies at every defensive line position.
--The Detroit Lions lost to the Washington Redskins three times between 1978 and 1981, and had 200+ rushing yards in all three losses.
--Nine quarterbacks have won two games in the same season, when their defense gave up 200+ rushing yards. Only Roger Staubach (1979) did it three times in the same season.
--Dan Marino is not one of those nine quarterbacks, though he is tied for the most overall since 1978, with five wins when his team gave up 200+ rushing yards.

We'll make this a trivia question: can you guess who else has won five games (1978 to present) in which his defense surrendered 200+ rushing yards?

Hints: It's not Peyton Manning (that was #4 for him on Monday night), and the quarterback in question is the only one to beat the same team twice in the same season, when his team gave up at least 200 rushing yards in both games. One of those games came in the playoffs.

The names of the 131 candidates on the preliminary list of modern-era nominees for election to the Pro Football Hall of Fame’s Class of 2010 have been released. There are 103 players on the list, including punters and kickers. In its current form, Approximate Value does not assign points to kickers and punters, although PFR favorite Nick Lowery is on the preliminary list. Of the remaining 99 players, they are presented below, sorted by career AV:

160	WR 		Jerry Rice
129	RB 		Emmitt Smith
121	QB 		Ken Anderson
114	DE/LB 		Chris Doleman
111	LB 		Rickey Jackson
107	DT 		John Randle
107	QB 		Randall Cunningham
105	QB 		Boomer Esiason
104	WR 		Tim Brown
104	CB/S 		Aeneas Williams
102	RB 		Ricky Watters
101	TE 		Shannon Sharpe
100	DT 		Steve McMichael
100	T 		Mike Kenn
100	LB 		Sam Mills
99	QB 		Rich Gannon
99	DT 		Cortez Kennedy
98	WR 		Cris Carter
98	LB 		Hardy Nickerson
98	CB 		Eric Allen
98	LB/DE 		Pat Swilling
97	DE 		Richard Dent
97	WR 		Andre Reed
97	LB/DE 		Kevin Greene
97	DE 		Too Tall Jones
96	WR 		Henry Ellard
96	T 		Lomas Brown
95	CB 		Louis Wright
93	LB 		Clay Matthews
93	RB 		Roger Craig
93	DT/DE 		Ray Childress
92	LB 		Karl Mecklenburg
91	QB 		Phil Simms
90	G 		Steve Wisniewski
90	S 		Donnie Shell
89	LB 		Cornelius Bennett
89	DE 		Clyde Simmons
89	DE 		Leslie O'Neal
88	S 		LeRoy Butler
87	QB 		Joe Theismann
87	WR 		Cliff Branch
86	WR 		Gary Clark
86	DT/DE 		Michael Dean Perry
85	QB 		Jim Plunkett
85	DE/LB 		Charles Haley
84	T 		Joe Jacoby
83	C 		Dermontti Dawson
83	G/T 		Chris Hinton
83	NT 		Fred Smerlas
83	S/CB 		Carnell Lake
82	S 		Steve Atwater
82	WR 		Stanley Morgan
81	WR 		Herman Moore
81	DE 		Charles Mann
81	CB 		Albert Lewis
80	RB 		Ottis Anderson
79	WR 		Irving Fryar
79	S 		Darren Woodson
78	RB 		Herschel Walker
77	CB 		Raymond Clayborn
77	T 		Jim Lachey
77	RB 		Eddie George
76	DE/DT/NT 	Joe Klecko
76	LB 		Matt Blair
75	G 		Max Montoya
75	DE 		Al Baker
75	LB 		Chris Spielman
75	S 		Joey Browner
75	LB 		Rod Martin
74	C 		Kent Hull
74	C 		Jay Hilgenberg
73	RB 		Terrell Davis
72	CB 		Lester Hayes
71	QB (also P) 	Danny White
70	LB 		Darryl Talley
68	WR 		Roy Green
68	TE 		Russ Francis
68	WR 		Sterling Sharpe
66	DE 		Dexter Manley
66	LB 		Ken Harvey
66	RB 		Larry Centers
65	CB 		Frank Minnifield
65	S 		Kenny Easley
64	TE 		Todd Christensen
63	T 		Tony Boselli
63	G 		Russ Grimm
63	RB/WR/KR/PR 	Eric Metcalf
61	RB 		Mike Pruitt
60	TE 		Frank Wycheck
60	TE 		Ben Coates
59	T 		Jimbo Covert
56	QB 		Doug Williams
56	TE 		Mark Bavaro
55	RB 		Gerald Riggs
46	RB 		George Rogers
45	T 		Korey Stringer
43	RB 		Leroy Hoard
39	RB/KR/PR 	Brian Mitchell
 7	ST/WR 		Steve Tasker

And below, the list of the non-players up for induction:

Administrator 	Bobby Beathard
Administrator 	Gil Brandt
Administrator 	Leo Carlin
Administrator 	Frank "Bucko" Kilroy
Administrator 	Art Rooney, Jr.
Administrator 	Ron Wolf
Administrator 	George Young
Coach 		Bill Arnsparger
Coach 		Don Coryell
Coach 		Tom Flores
Coach 		Jimmy Johnson
Coach 		Chuck Knox
Coach 		Buddy Parker
Coach 		Bum Phillips
Coach 		Dan Reeves
Coach 		Buddy Ryan
Coach 		Clark Shaughnessy
Commissioner 	Paul Tagliabue
NFL Films 	Ed Sabol
NFL Films 	Steve Sabol
Official 	Jim Tunney
Owner 		Jack Kent Cooke
Owner 		Ed DeBartolo, Jr.
Owner 		Ole Haugsrud
Owner 		Art Modell
Owner 		Carroll Rosenbloom
Owner  		Bud Adams, Jr.
Trainer 	Otho Davis
K 		Gary Anderson
K 		Nick Lowery
P 		Ray Guy
P 		Reggie Roby

Other links:

Greatest QBs Ever
Most Dominant RBs Ever
Greatest WRs Ever
One of many Jerry Rice posts by Doug
Top Ten Tight Ends of All Time
Art Monk = Shannon Sharpe?
Greatest Ks Ever
Greatest Coaching Records of All Time

We have 29 legal entries into the PFR season prediction contest. What was the best strategy? We will find out. If I told you who the three teams with the most wins would be, would you want to put them all in the first group, or spread them out? Well, if I told you, you would want to spread them out among the top groups, because you don't get bonus points for how much your group wins by. However, we don't know for certain, so when I run some simulations, putting those you think will do best (recognizing you are not perfect) is probably just as good as spreading them out.

Because of the multiplication of team wins, what you can't have happen if you want to do well is to have one of the three teams in your top groups get 5 or fewer wins, or two of your teams in the bottom groups do at least average. For that reason, I would have used the playoff wildcards earlier than most (to get a likely "9" but little downside), and coupled them with two strong teams. It will be interesting to see if the entries that used the wildcards earlier actually do perform better.

What I do know is that no two entries are alike, and you are mostly competing against your own entries. If you believe in the collective wisdom of the PFR readership (and if you want to assume that group placement typically translates to where people think a team will finish) then the playoffs would look like this:

AFC
1-New England
2-San Diego
3-Pittsburgh
4-Indianapolis
5-Baltimore
6-Tennessee

NFC
1-New York Giants
2-Minnesota
3-New Orleans
4-Arizona
5-Philadelphia
6-Green Bay

Here are the average group placement for each team and wildcard entry in the contest:

team	average
01-new	1.5
02-sdg	1.7
03-pit	2.0
04-clt	2.4
05-nyg	3.2
06-phi	3.5
07-PLA	3.5
08-min	3.8
09-nor	4.0
10-rav	4.2
11-gnb	4.5
12-chi	4.7
13-oti	4.8
14-PLN	4.9
15-atl	5.5
16-ari	6.0
17-dal	6.0
18-car	6.1
19-hou	6.1
20-sea	6.5
21-mia	6.8
22-jax	6.9
23-was	7.3
24-cin	8.3
25-nyj	8.9
26-HCA	9.1
27-sfo	9.1
28-HCN	9.6
29-ram	9.7
30-tam	9.9
31-buf	10.1
32-rai	10.1
33-cle	10.5
34-den	10.6
35-kan	10.6
36-det	11.4

Here is the variance of the picks, sorted from most variance to least. This should give you an idea of the biggest swing teams, where entries were in the most disagreement.

ari	8.6
gnb	7.0
jax	6.2
was	6.1
hou	5.6
sea	5.5
mia	5.4
chi	5.3
ram	5.3
oti	4.7
atl	4.1
dal	4.0
nyg	3.9
phi	3.7
buf	3.7
tam	3.6
nor	3.5
rav	3.5
nyj	3.3
sfo	3.3
car	3.2
cle	3.2
kan	3.1
rai	2.8
min	2.7
HCN	2.7
PLA	2.3
HCA	2.2
den	2.2
PLN	2.0
cin	2.0
clt	1.8
sdg	1.7
new	1.1
pit	0.9
det	0.5

With most of week 1 games in the books, I decided to play around with the P-F-R play index. Here's a quick recap of some of the queries I ran:

• Mark Sanchez became the first rookie QB to win a week 1 start and throw for over 250 yards since Fran Tarkenton.

• Sanchez (22 yrs, 306 days) is now the youngest QB to win a week 1 road game since 1960 (and maybe ever).

• If you really want to stack the deck to find some good Sanchez precedent, take a look at all week one rookies from 1961 to 2008, who at age 23 or younger, had 15 attempts and averaged at least 7.0 AY/A.

• Drew Brees' six TD passes were an opening day record: Ten times since 1960 has a QB thrown for exactly five TDs in week 1.

• Was it the best QB opener ever? Brees becomes the 10th QB to complete 75% of his passes, average 10 yards per attempt and throw 3 TDs in week one.

• Was it the best Romo opener ever? Tony Romo, without #81, becomes the first QB to throw for 3 TD, 0 INT, and average 13.0 yards per attempt in a season opener since... Tony Romo.

• All Day Peterson is now the 4th player to run for 150+ yards and 3+ TDs on opening day. Can you guess the first three?

• Yesterday was the 11th opening day performance of 50+ receiving yards for Isaac Bruce. We know who is #1, but can you guess who Bruce joined at#2?

• That guy in Minnesota tied Elway for most opening day wins ever by a QB with at least 5 pass attempts (a proxy for starts):

• It's okay, Jake. The last QB to get blown out and throw four picks on opening day won the SB that year.

• What a finish in Denver. Stokley/Orton tied for the longest fourth quarter touchdown by a team trailing by less than a TD that went on to win the game.

• Percy Harvin had over 20 rushing yards, 30 receiving yards, and caught a TD pass in his first professional game. [Edited thanks to Chris in comment 8.] That's the third straight year that's happened, after it hadn't happened for over a decade.

• Tim Hightower had 12 receptions against the 49ers. Pretty good, but not the most receptions in an opener by either a RB or a Cardinal.

• Reggie Wayne became just the second player since 1960 to have 10 receptions, 150 receiving yards and a score in an opening day victory. Do you know who the other is?

• After the game, Bart Scott said the Jets were swagger-licious. He's the first linebacker to change teams in the off-season and use that term following a week one road victory.

This year, we are going to do a team prediction contest. However, we can't just do a contest where you predict each team's record and who makes the playoffs. That would be too simple. Here's how it's going to work:

1) Each contestant will group the 32 NFL teams and 4 wildcard entries into 12 groups of 3 teams and rank order the groups;
2) The four wildcard entries are as follows:

(a) PLA (for playoffs, afc) will consist of the average 2009 win total of the 2008 AFC playoff teams (pittsburgh, baltimore, san diego, tennessee, indianapolis and miami)
(b) PLN (for playoffs, nfc) will consist of the average 2009 win total of the 2008 NFC playoff teams (arizona, philadelphia, carolina, new york giants, atlanta, minnesota)
(c) HCA (for head coach change, afc) will consist of the average 2009 win total of the AFC teams that have a different coach than the one who started the 2008 season (cleveland, denver, indianapolis, kansas city, new york jets, and oakland)
(d) HCN (for head coach change, nfc) will consist of the average 2009 win total of the NFC teams that have a different coach than the one who started the 2008 season (detroit, san fransisco, saint louis, seattle and tampa bay)

3) Only one wildcard entry may be used in each group of three. In other words, each group must have at least two NFL teams in it.
4) The three teams (or two teams and a wildcard) within each group will have their win total multiplied against each other. So, if you had the Giants, Eagles and AFC Playoff Teams (PLA) in a group, and they won 10, 8, and 9.5 games respectively, that would be a score of 760 for that group. Tie games in the NFL count as half-wins for the purpose of this contest.
5) Each contestant lists their 12 groups in descending order of how they think they will do under this scoring system.
6) A point is earned each time a group you have ranked higher outscores a lower ranked group. If two of the contestant's groups are tied with the same group points, a half point is earned. There is a maximum of 66 possible contest points (11 if your first ranked group outscores all others, 10 if your second ranked group outscores all those below, etc.)
7) In the event that there is a tie for most points among multiple entries, the first tiebreaker is most points scored by the first-ranked group, then most points scored by the second ranked group, and continuing to the third, fourth, etc., until the tie(s) is/are broken. In the event that two or more entries are still tied after the above step, the earliest entry wins.

List a three letter abbreviation for each team. I'll accept it as long as it logically makes sense and is not ambiguous as to what team is intended. We decide what is ambiguous. If you want to use "oti" or "ten" for tennessee, then that is fine. Let's go through a sample entry, using the PFR 3 team codes and the wildcard codes above, to see what an acceptable entry would look like.

1-mia/nyg/pit
2-min/oti/car
3-sdg/ari/pla
4-nwe/phi/pln
5-rav/chi/clt
6-atl/den/sfo
7-nyj/dal/gnb
8-cin/htx/hca
9-tam/sea/hcn
10-rai/was/buf
11-cle/det/nor
12-jac/kan/ram

So, in this example, the mia/nyg/pit group would get you a point for every other group it outscored. The rav/chi/clt group (group 5) would get you a point for every group between 6 and 12 that it outscored, and cost you the chance to score points if it outscored the groups above it.

All you have to do is make sure you have every NFL team listed once and only once, you use all four wildcard entries, and you have exactly three entries in each of twelve groups.

All entries must be in before 7 p.m. central time on Thursday, September 10, 2009.

Become a fan of PFR on Facebook

It is social media day on the Sports Reference sites.

I’m going to jump the gun a little here. I don’t have a strong opinion on whether the NFL should remain with a 16-game schedule or expand the regular season to 18 games. I don’t know what the right answer is here. There is a limit to how many games of football can be played in a year, I just don’t know if we are at that limit.

The one thing I do have an opinion on is that the league needs to do something to improve the competitive incentive for the final week of the season, so that it is not like a fourth pre-season game for a lot of teams. A year and a half ago, I wrote about a proposal to eliminate automatic home games for division winners now that we have four divisions of only four teams each in both conferences. A similar proposal was raised by the competition committee but failed to pass.

Despite my complaints about the uptick in competitive games in week 17 because wildcard teams have no chance at a home game, and division winners get locked in earlier with fewer teams in a division, the current scheduling format has several things going for it. First and foremost, it is simple and consistent. Compared to the scheduling formats used in the 1970’s and 1980’s, it is far fairer within divisions and less complex. Teams also get to play non-divisional opponents on a set rotating schedule, and these matchups are guaranteed—unlike the past where things like Miami going 15 years without playing a game in Denver were possible.

The rotating conference schedules are both good and bad, though. Good for simplicity and insuring that conference members will play each other regularly. Bad because it basically creates two separate groupings within a conference, where there are very few common games across groups. Yet, when tiebreakers like conference record come into play, completely different conference schedules, without head to head matchups, may be determining playoff spots and seeding.

If the league is going to add games, I have an alternate proposal to keep the current uniform 16-game schedule format, while creating exciting and competitive matchups in the final weeks with the additional two games.

Play flex games in the final two weeks of the regular season, setting the matchups based on how the season has progressed. The recent television deal with NBC on Sunday Night Football has already introduced the concept of a flexible schedule in terms of setting the night matchup closer to the time of the game. This idea simply builds on that. My idea would simply pair up those teams that have something to play for within a conference, and set up matchups that would decide playoff positioning on the field. For teams that had nothing to play for, the matchups could still be set up in such a way that geographic rivals can play at season’s end.

Here’s how my idea would work:

1) After 16 games have been played, the final two matchups are set based on record and playoff eligibility to that point.

2) The home team dates will be known ahead of time, by setting which divisions will have home games for game 17 versus game 18, so tickets for the games can be sold with just the identity of the opponent to be determined. For example, in 2009, the AFC West and North play each other. The AFC West/North would be at home for week 18 and on the road in week 19. The reverse would be true in the NFC, so that there would still be a roughly equal geographic distribution of home games both weekends.

3) All teams that have not clinched a specific playoff seed (even if they have clinched a playoff spot) and all remaining teams that are mathematically alive for a playoff spot are put into the Playoff Pool of teams.

4) All teams that have been eliminated from playoff contention, or have clinched a specific playoff seed are put into the Non-Playoff Pool of teams.

5) If the number of playoff pool teams within a conference is imbalanced across the two halves of a conference (In 2009, West+North versus East+South), then the necessary number of teams will move up from the Non-Playoff Pool to give an even number of matchups for Playoff Pool teams. The Non-Playoff Pool Team with the best record, from the half of the conference with fewer teams, will move to the Playoff Pool, if necessary, to balance out the matchups.

6) For Playoff Pool games, all games will be played between conference opponents. AFC Playoff eligible teams will only play AFC, and NFC teams will only play other NFC teams.

7) For Non-Playoff Pool teams, games can be played against both conference and non-conference opponents.

Teams will not play a non-divisional opponent they already played in the regular season in the flex games.

9) The matchups within the Playoff Pool will be set based on a priority order (subject to Rule #8). I would love to let teams select their home opponent, but I doubt this would ever happen in real life. Teams that were tied for a playoff position would meet, with the team holding the current tiebreaker getting home field for the matchup. The priority rules for setting matchups would have to be spelled out in detail, but for now, let’s say generally that first priority would go to teams that had not clinched a playoff spot, but were in playoff position after 16 games, followed by teams that had clinched playoff berth but not positioning, followed by teams out of playoff position but still in contention.

If we applied that to last year, here is one version of the matchups that could have resulted. The Ravens and Patriots ended up tied on record with the Patriots missing the playoffs on a tiebreaker. In this system, those teams meeting is a high priority, and New England at least gets to play their way in on the field in a play-in game in Week 19. In the NFC, several teams were still in playoff competition for that final spot, and the Eagles get to play two of the teams right behind them, while several other matchups also feature teams directly fighting for playoff spots.

Week 18
Baltimore at San Diego
Tennessee at New England
Pittsburgh at Denver
Indianapolis at Miami
San Fransisco at Atlanta
Washington at Minnesota
Dallas at Carolina
NY Giants at Tampa Bay
Arizona at Chicago
Philadelphia at New Orleans
Houston at NY Jets
Cincinnati at Buffalo
Saint Louis at Green Bay
Seattle at Oakland
Jacksonville at Kansas City
Cleveland at Detroit

Week 19
New England at Baltimore
San Diego at Tennessee
Denver at Indianapolis
Miami at Pittsburgh
Tampa Bay at Philadelphia
New Orleans at Arizona
Chicago at New York Giants
Minnesota at Dallas
Carolina at San Fransisco
Atlanta at Washington
Buffalo at Houston
NY Jets at Jacksonville
Green Bay at Cincinnati
Detroit at Seattle
Oakland at Cleveland
Kansas City at Saint Louis

Another trivia episode.

Doug and JKL both make complete fools of themselves early. JKL later redeems himself, while Doug goes on to make an even bigger fool of himself. Chase turns in a workmanlike effort. Play along, it's fun!

Listen here, subscribe here if you know how, and read this if you don’t. It’s free, of course.

The NFL is considering expanding the regular season to eighteen games, and one of the big points of discussion is what impact extending the regular season will have on injury risk to players. Brian Burke of Advanced NFL Stats and Bill Barnwell of Football Outsiders have already weighed in with some interesting thoughts on the topic, and the NFL’s study regarding injury rates.

Since it now seems somewhat relevant, I thought it would be as good of a time as any to dust off some data I had lying around from last off-season that never got published, and add in the new individual games info we have going back further than 1995 (as well as add in last season’s numbers).

I know the discussion on injury effects will center on more than the running back position, but it is this position, where the players take as many hits as any other and have shorter careers, that we probably all think of when we consider who may be most impacted by additional games.

I pulled all running backs who (1) had 200 or more rushes in a season from 1988-2007, and (2) played in the first game of the following season. This hopefully excluded all retirees and guys who came into the year injured. (On second thought, I'm extremely positive it did exclude all retirees). I then recorded which game, if any, each of the qualifying backs first missed a game that next season. So when you are looking at the chart below, the "Game 2" line tells you that 365 backs met the criteria by playing in game 1 the season after they had at least 200 rushes, 9 of those backs did not play in game 2, with 1 of them missing all remaining games starting with game 2. As you can see, the "Total" number is decreasing as the weeks progress because I am subtracting out the players who missed games. This is trying to look at the rate at which players who had not missed any games to that point then miss the next game. I should point out that not every game missed is due to injury, but I'm comfortable enough that the vast majority are missed games caused by injury.

GAME		Total	Missed	Missed All	Missed/Total
=====================================================================
2		365	9	1		0.025
3		356	21	3		0.059
4		335	19	2		0.057
5		316	14	3		0.044
6		302	23	3		0.076
7		279	10	3		0.036
8		269	11	2		0.041
9		258	14	3		0.054
10		244	5	1		0.020
11		239	10	3		0.042
12		229	13	5		0.057
13		216	11	3		0.051
14		205	8	4		0.039
15		197	7	5		0.036
16		190	14	14		0.074
play all	  176
=====================================================================

Judging by these numbers, there’s about a 4.7% chance that your established healthy running back who has not missed any games so far will then miss the next game, and a little less than 50% chance that a veteran back will play a full 16 games. I don’t see any particular trend that suggests that the injury rate is increasing in the final weeks of the season. You may be inclined to think it is, based on the percentage of guys who missed game 16. However, the group of backs who played 15 straight games then missed the final regular season game is over-populated with star backs on really good teams who had already clinched playoff spots. I don’t know individually how many would have still missed the final game if there was any incentive to play, but I feel pretty comfortable saying that the true numbers are a lot closer to the week 14 and 15 rates.

Will there be more injuries with two extra games? Sure there will, even if the rates don't increase. The powers that be on both sides of the negotiating table will have to decide what the appropriate level of risk is. Judging on these numbers, I don’t see that the rate of injury should be expected to increase dramatically in games 17 and 18. I will note, though, that this is looking only at the injury rate for guys who had not missed any games previously that same season. We could potentially have a situation where the running back injury rates are a little higher overall, even if the injury rates for previously healthy backs and injury rates for previously injured backs are steady, because there are more “previously injured” backs in game 16 than in game 4. (This assumes that players who have already missed games with injury and return within the same season are at an elevated risk of additional injury).

A bigger concern for me is not whether the additional games create more injuries in just those two weeks, but whether they increase the risk of serious injury going forward in the future. I’ve done research on injury rates in the past, including looking at end of season workloads and playoff games played.

That data was from 1995-2006 (I probably need to update that, but don’t have time right now). It looked at all players who averaged at least 15.0 carries a game over the final six regular season weeks and played in all games over that stretch. Pooling that data, the guys that didn’t play on playoff teams (n=104) suffered a season ending injury within the first 6 games of the following season 7.7% of the time. The playoff performers (n=86) had a season ending injury 7.0% of the time—roughly equal. The non-playoff guys missed at least one game within the first six weeks 19.2% of the time, versus 18.6% for the playoff guys. Average games played the following year was 13.5 for the non-playoff guys versus 13.3 for the playoff guys.

Now, if you look at the guys who played 3+ playoff games on top of the regular season, you might be concerned about increasing the game totals. I'll just note that of the 18 guys that averaged 15.0 or more carries and played in 3 or more post season games, the only 3 to suffer season ending injuries in the first month (and thus lower the games played dramatically for the group) were also the only 3 to average 25.0 or more carries a game over the final six weeks of the regular season. For backs with a more moderate starting running back workrate who were extended deep in the playoffs, the average games played the next year was 14.1.

From a reader named John Worrall, this appeared recently in my email box:

I have a theory that says that backup players that come into games perform at a very high level, because the opponent gameplanned for a different player. For example, when Joe Montana would go down with an injury and Steve Young came in, the opposing defense got a COMPLETELY different kind of QB than they were ready for, which meant Steve Young would have a huge game.

With all this new game log data we've got, we can test this kind of thing out.

The first thing I wanted to do was build a database query that would find all instances where one quarterback was unexpectedly replaced early in the game by another. I eventually decided to find all instances where:

1. the starting quarterback in a particular game threw five or fewer passes (and no interceptions), and the quarterback who replaced him threw 20 or more passes

and

2. the replacement started the following game, and the original starter (from the previous game) did not play.

From 1970 to 2008, there have been 73 such cases, so it's a respectable sample. Then I compared the replacement QB's adjusted yards per attempt ((yards + 20*TDs - 45*INTs)/attempts) in the first game, in which he was presumably not expected to play, to his adjusted yards per attempt in the second game, for which we can only assume the defense had a full week to prepare for him.

Here is a line of data that supports John's theory:

mia 1981  5 David Woodley        Don Strock            11.00   1.93

This says that in game 5 of the Dolphins' 1981 season, David Woodley started and Don Strock relieved him early in the game. Since Woodley had started all four of Miami's games before that one, the Dolphins were undefeated at the time, and nothing in the stat line indicates that Woodley did anything terrible to get himself pulled in that game, we have to assume that Strock's appearance was not expected by the Jets. Or the Dolphins, for that matter.

But anyway, Strock played very well, going 18 of 29 for 279 yards and two touchdowns. That's 11.00 adjusted yards per attempt, and you can click the '11.00' to see the boxscore of that game. In the following week, Strock started against the Bills and went 26 of 44 for 245 yards, with one touchdown pass and four interceptions. That comes out to 1.93 adjusted yards per attempt. 1.93 - 11.00 = -9.07, so this goes down as a -9.07, which represents strong evidence for the element-of-surprise-helps-backup-QBs hypothesis. Now all we have to do is repeat the procedure for all 87 other data points and average the numbers together.

Result:

On average, the replacement quarterbacks had about 0.15 more adjusted yards per attempt in the following game than they did in the original game. So the data seems to indicate that the quarterback himself benefits from the week of preparation as the starter at least as much as the defense benefits from preparing to face him. Interesting.

Just because I know someone will ask, I checked to make sure the two sets of defenses were comparable. They were. In the aggregate, the defenses faced by these QBs were actually negligibly better (.05 adjusted yards per attempt allowed during the given season) in the second games than in the original games.

Full data set below:

(Continued)

In the comments to Chase's post on the Jets' draft, football historian extraodinaire and p-f-r friend Sean Lahman pointed out that the Jets drafted Ken O'Brien, Al Toon, and Blair Thomas instead of Dan Marino, Emmitt Smith, and Jerry Rice. That prompted me to investigate other teams who might have just missed on all-time greats. It was so much fun that I decided to turn it into an eight-part series, looking at one division at a time. For no particular reason, I'll lead off with the NFC West.

Although it might make sense in some cases to look at the player(s) drafted immediately after a particular pick, I'm going to limit my attention to players at the same position. For example, in 1996, the Eagles took guard Jermane Mayberry one pick before the Ravens took Ray Lewis. Passing on Ray Lewis is never a good idea, but I'm not going to kill the Eagles for that pick. Instead of comparing Mayberry to Lewis, I'll compare Mayberry to the next guard taken, who happened to be the undistinguished Jason Layman. In that light, the Mayberry pick looks OK. I understand that that's an arbitrary choice, but it makes the comparisons a little cleaner. These are supposed to be fun posts, not serious analysis.

Before I get to it, let me first just say for the record that I understand that each team's draft board and draft strategy is different, and just because Layman was the next guard taken after Mayberry in no way means that the Eagles were sitting there on the clock trying to decide between Mayberry and Layman. So please understand that

Team X drafted Player Y instead of Player Z

and

Team X passed on Player Z to take Player Y

are merely shorthand for

Team X drafted Player Y, and the very next player taken at Player Y's position in that draft was Player Z

Again, this is just for fun.

Seahawks

In the big picture, Seahawk fans really don't have too much to complain about draft-wise, but Seattle's 1981 draft is perhaps one of the worst what-might-have-beens there is. And the killer is that it actually wasn't that bad of a draft. It's just that the guys they passed on turned out to be much better.

(Continued)

About a year and a half ago, I posted this batch of data on how teams are built. In it, I looked at where teams' 22 starters come from. But that's sub-optimal, because some starters are more important than others and also because backups have value too.

That post was written in the pre-AV era here at p-f-r. This weekend's draft inspired me to revisit it using the entire roster (weighted by AV) instead of just the starters.

I've broken things down by decade and by team quality (bad = 6 or fewer wins, mediocre = 7--9 wins, good = 10+ wins). The numbers in each category represent the aggregate percentage of the teams' AV that came from that category:

1980 -- 1989, bad teams
==============================
First round picks         23.6
Second round picks        14.9
Third/fourth round picks  19.3
Fifth round and later     28.7
Undrafted                 13.5

Drafted by team           80.5
Acquired via other means  19.5

Age 24 and younger        28.9
Age 25 to 27              37.8
Age 28 to 30              26.2
Age 31 and older           7.2

(Continued)

It took Art Monk eight years to make the Hall of Fame. While his career numbers were terrific, Monk's biggest problem was the lack of statistical single season dominance. He only ranked in the top 10 in receiving yards three times -- finishing fourth in '84, third in '85 and tenth in '89. But arguably Monk shouldn't have been compared to the star receivers of NFL history. As argued by Sean Lahman in the Pro Football Historical Abstract:

Even though Monk lined up as a wide receiver, his role was really more like that of a tight end. He used his physicality to catch passes. He went inside and over the middle most of the time. He was asked to block a lot. All of those things make him a different creature than the typical speed receiver.... His 940 career catches put him in the middle of a logjam of receivers, but he'd stand out among tight ends. His yards per catch look a lot better in that context as well.

I haven't heard anyone else suggesting that we consider Monk as a hybrid tight end, but coach Joe Gibbs hinted at it in an interview with Washington sportswriter Gary Fitzgerald:

"What has hurt Art -- and I believe should actually boost his credentials -- is that we asked him to block a lot," Gibbs said. "He was the inside portion of pass protection and we put him in instead of a big tight end or running back. He was a very tough, physical, big guy."

Monk has said similar things:

“In [1981] we were pass oriented and that didn’t work so well. So we went to a ground game. About this period of time we shifted a little into more of a balanced offense. I was moved from being just a wide receiver to playing H back. I would come out of the backfield and do a lot of motion. And we had a lot of success with that.”

More from Coach Gibbs:

'We used him almost as a tight end a lot,' said Gibbs, 'and not only did he do it willingly, he was a great blocker for us.'

It's an interesting argument, calling Monk a hybrid tight end. But now that Monk is in the HOF, the more interesting argument points to Shannon Sharpe. Is he the anti-Monk? While Monk may have been a tight end in wide receiver's clothing, was Sharpe a wide receiver in tight end's clothing? When Sharpe -- easily one of the greatest TEs of all time -- was not elected to the Hall of Fame in 2009, people were surprised. I believe Mike and Mike on the radio were the first to report this, but several claimed that the HOF voters were considering Sharpe as a wide receiver, and not as a tight end. My first reaction to this was probably like yours -- how ridiculous. But now I'm not so sure. Is Sharpe a hybrid wide receiver?

Baseball, Fantasy Football and the NFL

Jeff Kent was a very good hitter who played second base, a position that historically was played by good defensive players who were not great hitters. Kent may have been a 2B, but he was not very skilled defensively and may have been better suited at first base (which is where he ended his career). But by playing Kent at 2B, that allowed his teams to get another big bat into the lineup -- first base could be filled by a typical power hitter. If you put Kent at first base, you're going to put your typical lightweight hitting 2B into the lineup. So the trade-off is a slightly worse defense but a much better offense. As long as your power hitter is respectable on defense, putting him in at 2B instead of 1B makes your whole team better.

In fantasy football, this becomes even more obvious. In Marques Colston's rookie season he was listed as a TE in some fantasy leagues. Playing him at TE instead of WR left room for an extra WR -- a fantasy owner could play Colston and three WRs instead of Colston, two WRs and a TE. Since WRs score many more points than tight ends, this made him one of the most valuable players in fantasy football leagues.

But that's *not* the case in real football. And that's why the situations are apples and oranges. There's nothing magical about the name you give to a player's position. The Broncos used Terrell Davis (RB), Ed McCaffrey (WR), Rod Smith (WR), Howard Griffith (FB) and Shannon Sharpe. The Redskins used Earnest Byner (RB), Gary Clark (WR), Ricky Sanders (WR), Don Warren (TE) and Art Monk. Is there a meaningful difference between those lineups? With Sharpe, you still needed a guy like Griffith in there to have six blockers. With Monk, Gibbs was still able to get two other athletic wide receivers on the field.

When we think of a tight end, we think of a hybrid blocker-receiver. But, as Lahman says:

[In the 1990s,] the tight end began to re-emerge as a major part of the offense. Rather than look for a player who could both block and catch passes, most teams split the position into two roles. There were blocking tight ends and receiving tight ends, and two guys would replace each other as the situation dictated. What this meant was that a receiving tight end didn't need to carry the bulk necessary to block a 300-pound lineman, so the position could be stocked with smaller but stronger and more athletic players. The key figure in that last shift was Shannon Sharpe, who emerged as a new kind of offensive weapon with his play for the Broncos. He was just 6-foot-2 and 225 pounds, which most people considered too small to play tight end in the NFL. But he had the speed, strength and agility to create havoc for defenders.

During Sharpe's best years, the Broncos still carried two big blocking tight ends -- Byron Chamberlain (6'1, 250) and Dwayne Carswell (6-3, 290) on the roster. When Denver needed an extra blocker, those guys came in. And while Sharpe may have been a better blocker than guys like Dallas Clark or the Jets' Dustin Keller, these H-Back/slot receiver types are evidence that the tight end and the wide receiver positions are not binary options but rather they fall on a continuum. On one end, there are guys like Bob Hayes; closer to the middle are Art Monk and Shannon Sharpe and maybe a Hines Ward; on the other end is Bubba Franks. In that light, it's legitimate to wonder -- how much difference was there between Art Monk and Shannon Sharpe?

PFR lists Sharpe at 6'2, 225 and Monk at 6'3 and 210. While Sharpe looks a lot bigger, and their careers overlapped, some significant changes occurred in the NFL while these guys were playing. In Monk's breakout season, 1984, the average TE was 6'3 or 6'4 and 236 pounds. Ten years later, the average TE was 6'4 and 254 pounds. So Monk was about 25 pounds lighter than the average TE; Sharpe was a little shorter and about 30 pounds lighter than the typical tight end. In Monk's five 1,000 yard seasons, he averaged 13.8 yards per reception; the league average YPR for WRs was 15.2 in those seasons. In Sharpe's four big yardage years he averaged 13.0 YPR while the average WR averaged 13.7 yards per reception. Both were dependable, reliable possession receivers and had significantly better hands than the typical tight end. Both were much better blockers than your average WR but worse blockers than the average tight end.

If Sharpe is considered as a WR, he's in trouble. He ranked in the top ten just once in receiving yards, a tenth place finish in 1993. Like Monk, he has three Super Bowl rings, but that won't be enough if people compare him to Harrison, Owens and Moss. But the point of this post is that we shouldn't just think of these guys as tight ends or wide receivers, but as football players. And unlike in baseball, your contribution to your team can't be measured by what designation they put next to your name on the team roster.

In part one I attempted to adjust quarterback win-loss records for the differing levels of defensive support that each quarterback got. This post is an attempt to refine that idea.

Let me start by reiterating a couple of things I said in the previous post, but probably wasn't clear enough about.

Quarterbacks do not have win-loss records. Teams have win-loss records. Quarterbacks are big parts --- the single biggest individual part in most cases --- of teams, but they are not the same thing as teams. They therefore do not have win-loss records. Wins and losses are determined by (1) the quarterback, (2) the other offensive players, (3) the defense, and (4) the special teams (and also coaching, to the extent that it influences the three things above). So when people talk about "John Elway's record," they're really talking about something that was influenced by all four of the above factors in the games that Elway started.

The last post and this one are an effort to remove factors (3) and (4) as much as possible from the computation. What would "Elway's record" have been if he had average contributions from the defense and the special teams? (And, as discussed last time, understanding exactly how Elway himself might have contributed to (3) complicates things further.) These posts make no attempt to estimate what "Elway's record" would have been if he had had Jerry Rice to throw to in the 80s, or if he hadn't had Terrell Davis to hand off to in the 90s.

In order to keep this from reading like a software license agreement, though, I'm going to use terms like "Elway's record..." or "Elway won..." or "Elway was X games above average." These statements mean, respectively, "The teams quarterbacked by Elway had a record of ....", "the teams quarterbacked by Elway won....", and "The OFFENSES quarterbacked by Elway were X games above average."

And even with all those disclaimers out of the way, I'm not really sure how much I like the rankings from part one or the rankings I'm going to develop here. The only thing I'm sure of is that I like them better than regular win-loss records. I'm not making any claims beyond that.

OK, let's get on with it.

(Continued)