The Kansas City Chiefs had a regular season record of 47-33 over the last 5 seasons, tied with the San Diego Chargers for the most wins by a team that did not make at least one conference championship game appearance over that time. Many of the key pieces that contributed to those wins are no longer in place heading into next season. Trent Green was recently traded to Miami. Future Hall of Famer Will Shields retired in the offseason, one year after future Hall of Famer Willie Roaf retired. Priest Holmes has faded away without officially retiring as of yet, replaced by Larry Johnson. The Chiefs have transitioned from Dick Vermeil to Herm Edwards, and it appears as though the previous era has come to a close.

I am going to look at the prospects for franchises who had similar periods of some decent regular season success, but did not advance far into the playoffs over the entire period. These are teams, like Kansas City, that were on a plateau of perpetual promise. I searched each franchise starting in 1966 and moving forward in time, and set the following parameters to identify similar "plateau" teams from the past:

• The team won at least 50% of its games over a five-year period;
• The team never had a record worse than 6-10 (or its equivalent) in any single season, and never had a record worse than 14-18 (or its equivalent) in any two consecutive years within the five-year period;
• The team did not advance to a conference championship game in any of the five years in question; and
• The team did not win two-thirds or more of its games in the final year (year 5). (This is to avoid catching emerging teams, such as 1996 Denver).

I moved forward chronologically until a five-year period for a franchise met all the criteria. Some franchises continued to meet the criteria for a plateau team in additional seasons. I did not include those unless they involved a completely independent five-year period. For example, New Orleans 1986-1990 and New Orleans 1991-1995 are both included; New Orleans 1988-1992 is not.

Kansas City 2001-2005 met all the criteria, along with 38 other teams since 1966. Some of these plateau teams were directly on the heels of a more successful period--13 of these "plateau" periods immediately followed a season where the franchise had been to a championship game or beyond. Others emerged from losing seasons to show promise, but then stall. Still others, like New Orleans of the early 1990's, Philadelphia of the mid-1990's, and Miami from the start of this decade, were seemless continuations of previous plateau periods.

Here is the list of the "plateau of perpetual promise" teams, listed in chronological order, with the five year record for the team also listed.

team   years          record
=======================================
sdg  1966-1970      38-27-5
ram  1967-1971      49-16-5
det  1969-1973      40-26-4
cin  1973-1977      46-24
was  1973-1977      47-23
crd  1974-1978      44-28
mia  1975-1979      47-27
nwe  1976-1980      50-26
chi  1976-1980      40-36
atl  1977-1981      41-37
min  1978-1982      36-36-1
buf  1979-1983      40-33
gnb  1981-1985      37-35-1
cin  1982-1986      39-34
dal  1983-1987      45-34
nyj  1983-1987      41-38
sea  1984-1988      48-31
nor  1986-1990      46-33
phi  1987-1991      48-31
mia  1987-1991      42-37
oti  1988-1992      49-31
pit  1989-1993      45-35
min  1989-1993      44-36
rai  1991-1995      43-37
nor  1991-1995      45-35
phi  1992-1996      46-34
mia  1993-1997      45-35
buf  1994-1998      43-37
kan  1994-1998      51-29
sea  1995-1999      40-40
was  1996-2000      41-38-1
mia  1998-2002      50-30
den  1999-2003      44-36
nyj  1999-2003      42-38
gnb  1999-2003      51-29
nor  2000-2004      42-38
bal  2001-2005      42-38
kan  2001-2005      44-36
ram  2002-2006      41-39
===============================================

Let's start with the bad news. Here are the next three seasons post-plateau period. "Bad" stands for the percentage of teams that won one-third or fewer of their games in a season (5-11 in a 16-game schedule). "Good" stands for the percentage of teams that won at least two-thirds of their games in a season (11-5 or better). "Play" is for playoff appearances, "Champ" is for Championship Game Appearances, "SB" is for Super Bowl Appearances, and "WIN" is for Super Bowl Wins.

YR   NO.     BAD  GOOD  PLAY  CHAMP  SB    WIN      
===============================================
1    38     .237  .184  .368  .053  .000  .000
2    36     .389  .167  .250  .111  .056  .000
3    35     .200  .086  .200  .057  .029  .000
===============================================

For comparison, taking into account when these teams were playing, and the various playoff structures and league sizes, a randomly selected "average" team should make the playoffs about 38% of the time, reach a championship game 14% of the time, and go to a Super Bowl in 7% of the seasons.

Now, for the good news, for those with patience to endure short term set backs. Here are Years 4-8 post-plateau period:

YR   NO.     BAD  GOOD  PLAY  CHAMP  SB    WIN      
===============================================
4    32     .094  .219  .469  .125  .031  .000
5    31     .097  .387  .548  .355  .161  .097
6    31     .194  .290  .419  .161  .097  .032
7    30     .200  .300  .400  .233  .100  .000
8    29     .241  .207  .345  .207  .138  .034
===============================================

Year 5 post-plateau was a particular good year, and several teams that were emerging by years 4 and 5 continued to be dominant forces for several years. This includes the 1991-1995 Dallas Cowboys, the 1982-1985 Washington Redskins, the 1984-1988 Chicago Bears, the 1988-1992 Buffalo Bills, and the 2001-2004 Philadelphia Eagles. The teams that were re-emerging four to five years later were generally doing so with different stars. In fact, Kenny Anderson of the 1977 and 1981 Bengals was the only quarterback who was the team's leading passer at the end of a plateau period, and again when the team advanced to a championship game four to five years later. Walter Payton (1980 and 1984-1985), was the only running back who was leading rusher at the end of a plateau and again when the team reached a championship game. Only two receivers accomplished the feat--Stanley Morgan of the Patriots (1980 and 1985), and Tim Brown (1995 and 2000).

When we sort these plateau teams by what they did in the first two years after the plateau period, something very interesting emerges. Kicking out the teams that concluded their plateau period since 2002, I sorted the 31 remaining teams into three roughly equal groups. The bottom group I will call the "Two Steps Back" group, consisting of teams that won 11 or fewer games in the two years following the plateau period. The middle group is "Mired in Mediocrity", and were teams that won between 12 and 17 the next two years. Finally, the upper group is the "Reloaded" group, all with 18 or more wins. The numbers below are the combined numbers for years 4-8 post-plateau for each group.

GROUP    NO./Seasons    BAD  GOOD  PLAY  CHAMP  SB  WIN      
=======================================================
TwoSteps    11/52     .096  .462  .615  .365  .192  .077
Mediocrity  10/47     .298  .085  .234  .043  .043  .000
Reloaded    10/50     .120  .300  .480  .240  .080  .020
=======================================================

The Teams that went "two steps back" in the short term really did make the best long-term choice. Most of these teams that stepped back made adept choices with their early draft picks, and laid the foundation for future success. Cincinnati added Anthony Munoz in 1980. Between 1980 and 1983, the Bears drafted LB Otis Wilson, OT Keith Van Horne, QB Jim McMahon, OT Jimbo Covert, and WR Willie Gault in the first round. The Bills added DE Bruce Smith, QB Jim Kelly, OT Will Wolford and RB Thurman Thomas during their "two steps back" (for the Bills, it was actually three years of 8 total wins). The Cowboys drafted Aikman, Irvin and Smith. The Eagles, now famously, selected Donovan McNabb rather than crowd choice Ricky Williams. Seven of the eleven teams who took "two steps back" reached a Super Bowl by Year 8 post-plateau. Almost half of the seasons from years 4-8 were "good" seasons, as defined above.

The "Reloaded" teams that brought in new key offensive personnel while continuing to win also fared pretty well, continuing to post a nice "good" season to "bad" season ratio, and making (and advancing) in the playoffs at an above average rate. 1982 Washington was the only team to win a Super Bowl. Several other "reloaded" teams reached championship games and Super Bowls, including the Rams teams' of the mid-1970's that lost 4 championship games before making the 1979 Super Bowl, the early Marino-led Dolphins teams, and the Steelers after Cowher took over in the early 1990's. Of the more recent plateau teams, Denver (10-6 in 2004 followed by 13-3 and champ game appearance in 2005, the Broncos transitioned to Cutler at QB and WR Javon Walker, and still posted a winning record last year) and Baltimore (13-3 last year) would fall in the "reloaded" category.

The "Mired in Mediocrity" teams, on the other hand, died a slow death. These teams did not maintain the same level of play as during the plateau period, but they did not completely collapse over the course of the next two seasons either. The two "mediocrity" teams that actually reached a championship game and beyond are the exceptions that prove the rule.

Seattle 2000-2001 won 15 games. They did so by going 6-10 in 2000 with Kitna, Watters, and Dawkins as leading passer, rusher, and reciever, and then turning things around to 9-7 in 2001, led by new pieces that would be around for the Super Bowl appearance, in Hasselbeck, Alexander, and Jackson. Thus, they did turn over key offensive personnel and get younger, but the record was better in year one than most teams that adopted that strategy.

The Houston Oilers 1993-1994, soon to be the Tennessee Oilers and then the Tennessee Titans, were the other "mediocrity" team to reach a championship game. They did so with 14 total wins in the next two years. In year 1, they went 12-4, had a first round bye, and lost to Kansas City at home. The Oilers then blew things up, and fell all the way to 2-14 the next year. But that giant step backward allowed them to get into the top of the draft and select Steve McNair, the franchise quarterback who would be leading them to the Super Bowl in 1999.

Four teams that finished their plateau period since 2002 qualify as "mired in mediocrity" teams based upon their win totals in the two years that followed. Those teams are 1998-2002 Miami, 1999-2003 New York Jets, 1999-2003 Green Bay, and 2000-2004 New Orleans. Personally, I think New Orleans is most likely to prove to be an exception, despite the fact that the last two years gives them 13 wins, technically placing them at the lower end of the "mired in mediocrity" group. It is rare that a team lands a pro bowl quarterback in his prime, while also finding the running back and receiver of the future in the same season. New Orleans more fits the profile of a "two steps back" team, but one who did not have to go through the growing pains of developing a franchise quarterback in year 2.

The Jets and Green Bay could go either way, though both still have the same quarterbacks in place, and I am doubtful that Favre will still be the QB when the Packers are ready to legitimately contend for the Super Bowl again. Miami is following the pattern of most previous mediocrity teams, unwilling to endure a longer term rebuilding project that could result in a better product in 3-4 years. Thus, entering year 5, the Dolphins have still not found the QB that will lead them to upper level playoff contention, choosing to go with Culpepper last year, and now an aging Trent Green this season.

Turning back to Kansas City, GM Carl Peterson has, until now, been unwilling to embark on a rebuilding project. His last opportunity was following the 1998 season, when Kansas City rebuilt the Vermeil era Chiefs offense through veteran free agents (Holmes, Kennison) and trades (Trent Green, Willie Roaf). There are signs that the Chiefs may be altering their approach this time around, but the signs are mixed. If the Chiefs study their history, the decision should be easy. This Chiefs team has only one guaranteed starter on offense (Larry Johnson) and one guaranteed starter on defense (Derrick Johnson) who were first day selections in the draft between 2001-2005. Thus, it is hard to imagine an immediate "reload" with so little production coming from those drafts that should be forming the core of the roster at this point.

Two decisions will demonstrate the Chiefs' choice. Starting quarterback is the first one. Huard is the safer choice if you want 7-9 wins, but he is not the long term answer. Croyle may or may not be the long term answer, but the Chiefs need to find out, at the possible expense of wins, so they know whether they will need to draft a quarterback the following season. The handling of Larry Johnson's contract demands is the other. Larry Johnson makes the Chiefs a better team in the short term. However, he is not likely to be productive when/if the Chiefs are ready to legitimately compete for championships.

If the Chiefs want to avoid going the way of the Falcons of the mid-1980's, the Jets and Seahawks of the early 1990's, and the Redskins since 2000, they need to make decisions that put them at risk of losing games in the short term, to find out what they have in place for the long term, and what they will need to add over the next two drafts. Of course, you still have to draft the right players if you are selecting early--something prior plateau teams have shown very capable of doing right after the plateau ends.

Get rid of the human polls.

As I've said before, "The BCS" has become one of those abstract entities, like "bureaucrats" or "the government", that essentially means "something I don't like." So many people dislike so many different things about the BCS that it's almost impossible to have a conversation about it. In my opinion, the funniest kind of anti-BCS rant is the one that rails against the computer rankings. Computers don't understand football! Computers can't weight the myriad strengths and weaknesses of each team!! A computer has never felt the raw intensity of a Saturday night SEC game!!! How can it possibly tell us whether one team is better than another?

Those are legitimate points. As it stands now, though, the alternative --- the human polls --- is even more rote, mechanical, and formulaic, and even less likely to have involved serious thought. In fact, I'm pretty sure that the human polls actually are computer algorithms, and the code looks something like this.

Rankings = LastYearsRankings
For i = 1 to NumberOfWeeks  {
  For j = 1 to NumberOfTeams  {
    if (Team j lost a close game in week i OR Team j lost to a better team in week i)
      drop team j four slots in the rankings
    else 
      drop team j nine slots in the rankings
    }
  move everyone who didn't lose up to fill in the gaps
  }  

I'm going to test this out in 2007 by seeing just how close this ridiculously naive system can get to mimicking the final human poll.

I'll start by taking last year's final poll*, and then filling in the unranked teams in order of their records, with ties broken by perceived strength of conference (SEC > Big 10 > Big 12 > Pac 10 > ACC > Big East). I'll define a "close game" as one decided by less than 10 points and a "better team" as one that was ranked higher (according to this system) at the time of the game.

* - In order to account for the perceived divide between the BCS conferences and the non-BCS conferences, I will rank all the BCS teams from top to bottom, then I will rank the non-BCS teams from top to bottom. Then, for the initial rankings, I will place all the BCS teams ahead of all the non-BCS teams. I think this is about right. The pollsters will allow a non-BCS team to get pretty high up in the rankings if they keep winning. But they will always choose a BCS team over a non-BCS team with an equal record.

Final note: I reserve the right to make changes to my algorithm until the season starts on August 30th, but the system will maintain the ridiculous naivety of the one I'm printing here. Feel free to add suggestions if you have them.

Every year during training camp, players proclaim to the world their lofty and unlikely-to-be-attained goals. Jon Kitna got an early start this year, predicting double digit victories for the Lions this season. Well in 2001, Curtis Martin had this to say:

"I do a running test now and don't even sweat," he said. "Last year, it was killing me. I'm stronger. I'm lifting more weights. I'm putting up more weight. I feel better year to year. I feel that I'm in the second half of my career. My goal is that my second six years is better than my first six years. That's my goal. I feel like I'll accomplish that."

Curtis Martin, at age 28, had rushed for 7,754 yards through his first six seasons. At that time, only five RBs had rushed for more yards than Martin through six years. Only one RB had more carries after six years than Martin, and he (Eric Dickerson) would go on to have one of the worst first half/second half splits you'll ever see. Martin's heavy workload was cited as one of the reasons he was banged up for most of 2000 and used to explain why his productivity had begun to decline. Many thought his best days were behind him and that he was washed up. And of course, only one RB -- Emmitt Smith -- has ever rushed for 7,754 yards (or more) after turning 28 years old, and he hadn't even done that at the time Martin made his remark.

All those thoughts went through my head six years ago when I first read Martin's quote. I thought he was crazy. Six years later, how did Martin do? While he would only play 11 seasons in his career, Martin managed to do a pretty darn good job of holding up. Consider:

                    Rush Yds/Gm   YPC    Games missed due to injury

First six years:       84.3       3.86   4

Last five years:       83.5       4.21   4

Twenty-six RBs rushed for 5,000 or more yards during the first halves of their careers, and only two -- Barry Sanders and Jim Brown -- actually ran for more yards in the second half on their careers. Here's the full list for all inactive running backs. For RBs that played an even number of seasons, the first half/second half split is straight forward. For RBs that played an odd number of seasons, the middle season is equally split between the first and last part of the player's career. So when O.J. Simpson rushed for 1,125 yards in his sixth of 11 seasons, the first and last halves of his career are each credited with 562.5 yards. Because yards just look weird with a decimal point in them, I rounded up.

Name            First   Last     Diff

Barry Sanders	6789	8480	-1691

Jim Brown	5759	6553	- 794

Curtis Martin	7152	6949	  203

O.J. Simpson	5744	5493	  251

Ricky Watters	5524	5119	  405

Marshall Faulk	6701	5578	 1123

Corey Dillon	6209	5032	 1177

Walter Payton	8997	7729	 1268

Tony Dorsett	7015	5724	 1291

Franco Harris	6836	5284	 1552

Eddie George	6120	4322	 1798

Jerome Bettis	7918	5745	 2173

Joe Perry	5337	3041	 2296

Terry Allen	5457	3157	 2300

Marcus Allen	7275	4968	 2307

Gerald Riggs	5268	2920	 2348

Freeman McNeil	5320	2754	 2566

Emmitt Smith	10697	7658	 3039

Herschel Walker	5652	2573	 3079

Terrell Davis	5409	2198	 3211

Earl Campbell	6457	2950	 3507

Calvin Hill	5009	1074	 3935

Thurman Thomas	8178	3897	 4281

Greg Pruitt	5022	 650	 4372

Eric Dickerson	9086	4174	 4912

Ottis Anderson	7843	2430	 5413

While lots of RBs fade into the sunset as they grow older, we all know that some RBs have had slow starts to their careers. Here's a list of all the inactive RBs to rush for over 1,000 more yards in the second half of their careers:

Name           		First   Last     Diff

Tiki Barber		2806	7642	-4836

Robert Smith		1829	4989	-3160

Rocky Bleier		 693	3172	-2479

Hewritt Dixon		 407	2683	-2276

Priest Holmes		2880	5156	-2276

Jim Otis		1053	3297	-2244

John Riggins		4655	6697	-2042

Earl Ferrell		 464	2486	-2022

James Stewart		2020	3821	-1801

John Henry Johnson	2507	4297	-1790

Barry Sanders		6789	8480	-1691

Joe Morris		2039	3546	-1507

Gaston Green		 321	1816	-1495

Charlie Harraway	 788	2231	-1443

Moe Williams		 195	1631	-1436

Tank Younger		1123	2517	-1394

Lamar Smith		1743	3106	-1363

Charlie Garner		2876	4222	-1346

Tom Matte		1652	2994	-1342

Stump Mitchell		1674	2975	-1301

Robb Riddick		  50	1291	-1241

Garrison Hearst		3369	4597	-1228

Charles White		 942	2133	-1191

Willie Ellison		1137	2289	-1152

Cannonball Butler	 880	1888	-1008

Warrick Dunn and Fred Taylor are two active players that might one day land on this list; both would make it if they retired today. Reuben Droughns is currently at -3133 and Thomas Jones is at -2229, and a host of younger players are in the high negatives as well.

What about on the other side? Here's the complete list of inactive players that had much better starts to their career than finishes:

Name           		First   Last    Diff

Ottis Anderson		7843	2430	5413

Eric Dickerson		9086	4174	4912

Greg Pruitt		5022	 650	4372

Thurman Thomas		8178	3897	4281

Calvin Hill		5009	1074	3935

Earl Campbell		6457	2950	3507

Ollie Matson		4194	 979	3215

Terrell Davis		5409	2198	3211

Herschel Walker		5652	2573	3079

Emmitt Smith	       10697	7658	3039

Rick Casares		4414	1383	3031

Joe Cribbs		4046	1310	2736

James Jones		3138	 488	2650

Freeman McNeil		5320	2754	2566

Larry Brown		4177	1698	2479

Billy Cannon		2449	   6	2443

Paul Lowe		3689	1307	2382

James Wilder		4178	1830	2348

Gerald Riggs		5268	2920	2348

Rueben Mayes		2898	 586	2312

Marcus Allen		7275	4968	2307

Terry Allen		5457	3157	2300

Joe Perry		5337	3041	2296

John Brockington	3718	1468	2250

Sherman Smith		2880	 640	2240

John Stephens		2809	 631	2178

Jerome Bettis		7918	5745	2173

Boobie Clark		2565	 467	2098

Tommy Mason		3135	1069	2066

Keith Byars		2584	 525	2059

Tony Galbreath		3063	1009	2054

Hugh McElhenny		3649	1633	2016

Mike Alstott (2196) and Anthony Thomas (2054) would join the list if they retired today. Earl Campbell may be lower than you'd expect, but he did have one 1300 yard season at the beginning of the second half of his career. Dickerson's a victim of having so many huge years early in his career; he still managed over 4,000 yards in the second half of his career. Ottis Anderson stuck around forever, but earned a memorable Super Bowl MVP award ten years after making his last Pro Bowl. A few more random thoughts...

• Floyd Little (3,162 yards), Jerry Latin (280), Mike Green (71), Scott Williams (37) and Steve Hendrickson (3) are the only RBs to finish with the same number of rushing yards in the first and second halves of their careers (excluding the lengthy list of RBs with zero career rushing yards). Willie Brown (133), Keith Kinderman (111), Kenny Gamble (24) and Leo Hayden (11) each played three years, recorded rushing yards (in parentheses) in the middle year, but did not record a rushing yard in their first or third season in the league. Reggie Branch pulled the same donut but over five years.

• Of all RBs that were not active in 2006, 365 of them rushed for more yards in the second half of their careers than in the first; 829 RBs rushed for more yardage in the first half of their careers.

• John Riggins, James Brooks, Garrison Hearst and Ricky Williams all rushed for over 3,000 yards in the first halves of their career, but rushed for more yards in the second half. Barry Sanders, Tiki Barber, Jim Brown, Riggins and Priest Holmes are the only RBs with over 5,000 rushing yards in the second halves of their careers.

• Be careful what assumptions you make from all of this. Just because a RB rushed for a lot more yards in the first half of his career than the second doesn't mean he broke down. If Thurman Thomas retired after an unimpressive 1997, he would have rushed for a little more than 1,000 more yards in the first half of his career than the second half. But because he stuck around (a sign that he didn't break down entirely), his split ends up looking much more heavily weighted towards the early part of his career.

The Eagles' decision to draft University of Houston quarterback Kevin Kolb with the 36th overall pick has cast some doubt on Donovan McNabb's future with the team.

For a historical perspective, let's look at all teams since 1978 that:

1. had the same primary quarterback for three consecutive years;
2. drafted a quarterback with a pick between #15 and #45.

The age listed is the incumbent quarterback's age at the end of the season just prior to the drafting of the young quarterback. For reference, McNabb was 30 at the end of last season.

Tm   Yr   Incumbent          age  Pick  Rookie QB
==========================================================
kan 1979 Mike Livingston      33   23 Steve Fuller
oak 1980 Ken Stabler          34   15 Marc Wilson
pit 1980 Terry Bradshaw       31   28 Mark Malone
buf 1980 Joe Ferguson         29   37 Gene Bradley
stl 1981 Jim Hart             36   33 Neil Lomax
nyj 1983 Richard Todd         29   24 Ken O'Brien
mia 1983 David Woodley        24   27 Dan Marino
cin 1984 Ken Anderson         34   38 Boomer Esiason
phi 1985 Ron Jaworski         33   37 Randall Cunningham
sea 1991 Dave Krieg           32   16 Dan McGwire
atl 1991 Chris Miller         25   33 Brett Favre
nyj 1991 Ken O'Brien          30   34 Browning Nagle
den 1992 John Elway           31   25 Tommy Maddox
kan 1992 Steve Deberg         37   40 Matt Blundin
buf 1995 Jim Kelly            34   45 Todd Collins
sfo 1997 Steve Young          35   26 Jim Druckenmiller
gnb 2005 Brett Favre          35   24 Aaron Rodgers

As we would expect, many of the incumbent quarterbacks --- Stabler, Hart, Anderson, Jaworski, Deberg, Kelly, Young, and Favre --- were older than McNabb and clearly in need of an heir to the job. Some of the others --- Livingston, Todd, and Woodley --- despite having been their team's quarterback for at least three years, simply weren't considered very good. O'Brien and Miller probably belong in this class as well.

That leaves Bradshaw, Ferguson, Krieg, and Elway. Although it clearly wasn't because Dan McGwire was too good to leave on the bench, Krieg played only one more season in Seattle. The other three all played at least three more seasons, and none ceded the job to the youngster drafted to replace him.

So if nothing else, we've learned that this situation is rare; it's been about fifteen years since we've seen something comparable. The Eagles shouldn't be too shocked that McNabb feels blindsided. At the same time, though, McNabb needs to realize that he --- not Kevin Kolb --- is in control of his future. If McNabb continues to play well, then he, like Bradshaw, Ferguson, and Elway, should continue to play for his current team for several more seasons.

This is a guest post by David Shick! Its purpose is three-fold:

1. To pass the time until the actual training camp news starts flowing;
2. to appreciate the under-appreciated career of John L. Williams;
3. to use the power of the p-f-r blog readership to determine whether Williams was, in fact, the last of his kind.

David's post is inside the quote-mark thingies. I've got a few comments at the end.

Head coach (Ground) Chuck Knox brought his run happy offense to the northwest in 1983. The Seahawks quickly traded up in the first round of Knox's first draft to take running back Curt Warner from Penn State University so Knox would have his featured tailback. Warner set Seahawk records for carries (335), rushing yards (1449), and rushing touchdowns (13) in his rookie campaign, but Knox wasn't happy. He wanted a fullback to complement his new weapon. To top that off Knox must have really blown a gasket when Warner blew his ACL only one game into the 1984 season. Dave Krieg attempted 480 passes that season. This would not do. With the first Seahawk pick in 1985 Knox selected fullback Owen Gill. Gill didn't make it out of training camp on the roster. Krieg attempted 532 passes in 1985. Knox was losing his mind. With the fifteenth pick in the first round of the 1986 draft Knox selected fullback John L. Williams from the University of Florida. Can you imagine your current favorite team selecting a fullback with their first overall pick in back to back years?

Williams had an immediate impact on the Seattle offense. He and starting tailback Curt Warner were both on the field for virtually every down. His career statistics tell the whole story. We're talking about a fullback here. Not a featured tailback. Warner never came close to the number of touches he had during his rookie year. He and Williams literally shared the load. During the bulk of his productive years in Seattle Williams was getting 10+ carries and 4+ receptions each game.

                 +--------------------------+-------------------------+
                 |          Rushing         |        Receiving        |
+----------+-----+--------------------------+-------------------------+
| Year  TM |   G |   Att  Yards    Y/A   TD |   Rec  Yards   Y/R   TD |
+----------+-----+--------------------------+-------------------------+
| 1986 sea |  16 |   129    538    4.2    0 |    33    219   6.6    0 |
| 1987 sea |  12 |   113    500    4.4    1 |    38    420  11.1    3 |
| 1988 sea |  16 |   189    877    4.6    4 |    58    651  11.2    3 |
| 1989 sea |  15 |   146    499    3.4    1 |    76    657   8.6    6 |
| 1990 sea |  16 |   187    714    3.8    3 |    73    699   9.6    0 |
| 1991 sea |  16 |   188    741    3.9    4 |    61    499   8.2    1 |
| 1992 sea |  16 |   114    339    3.0    1 |    74    556   7.5    2 |
| 1993 sea |  16 |    82    371    4.5    3 |    58    450   7.8    1 |
| 1994 pit |  15 |    68    317    4.7    1 |    51    378   7.4    2 |
| 1995 pit |  11 |    29    110    3.8    0 |    24    127   5.3    1 |
+----------+-----+--------------------------+-------------------------+
|  TOTAL   | 149 |  1245   5006    4.0   18 |   546   4656   8.5   19 |
+----------+-----+--------------------------+-------------------------+

Seattle fans loved Williams. He was our special weapon and no one else had one. Perhaps some of the softest hands out of the backfield in the history of the NFL. He caught everything remotely close. In interviews he was humble and soft spoken very similarly to Seattle Mariner DH Edgar Martinez. Produce on the field and finish it off with an "Aw, shucks fellas". My favorite Williams play was from a game in 1986. It was week seven at home against the eventual Super Bowl champion New York Giants. Late in the game with a one point lead Williams got together on the sidelines with QB Dave Krieg and head coach Chuck Knox. He described a problem with the Giant defense that lead to a vulnerability. If they ran a screen in just such a way he would easily get free. It led to a huge gain (fifty yards? I can't remember the exact yardage) setting up a short Warner touchdown sealing the win. That was the last game the 1986 Giants lost before posting twelve consecutive wins including a Super Bowl victory over Denver.

I was explaining Williams' uniqueness to my wife probing for another player similar to Williams in the past twenty years. Her only guess was Mike Alstott, but I wonder if he really counts. Was Alstott a fullback, or just an oversized tailback? I don't recall seeing him in games as a lead blocker on first and second down. The modern day fullbacks like Seattle's Mack Strong and San Diego's Lorenzo Neal rarely get touches. I would love to get some feedback on this concept. I ask because I don't know. Was the Warner/Williams combo really the last two back tandem that was on the field together every down sharing duties? Kiick and Morris? Bleier and Harris? I expect to hear about Bush and McAllister in New Orleans, but again, they're both tailbacks splitting carries. Neither is a real fullback, right?

For what it's worth, here are a few names to start the discussion:

• Williams' college teammate Neal Anderson was a similar player, and just as good. He eventually transitioned to a more traditional tailback role, but he shared the backfield early in his career with Walter Payton.
• After the aforementioned transition, Brad Muster shared the backfield with Neal Anderson. Muster wasn't a bad player, but he was no Williams.
• It seems likely that the Williams role is what the Browns had in mind for Tommy Vardell when they drafted him in 1992. That obviously didn't work out.
• This is pure speculation on my part, but it's conceivable that William Floyd might have become a Williams-like player had he not gotten hurt early in his career.

Prompted by the observation that this year's AFC West will likely feature a lot of very inexperienced quarterbacks I posted a list last Thursday of divisions that were collectively the least experienced at quarterback.

Commenter Vince suggested that it might be fun to look forward instead of looking back. In other words, rather than seeing how much previous experience the players had at the time, we could look at how much experience they would accrue after that point. So a division that scores high in this measurement is one that not only has good players, but good players near the start of their careers.

For example, look at the division that this year's AFC West will likely displace as the least experienced of the 16-game-schedule era: the 1985 AFC Central. Fans didn't know it at the time, but they were witnessing the early stages of some very good long careers. That division featured three first-year starters --- Bernie Kosar, Boomer Esiason, and Mark Malone --- and one second-year starter: Warren Moon. As you know, those players went on to play a lot of games from 1985 on. In fact, they threw more than 16,000 combined passes from that season on, an average of 4027 each. That's the second-highest figure for any division since 1978. Here are the top few. The numbers shown are pass attempts before that season, and passing attempts in that season and after.

1992 NFC Central       853   4191
Rich Gannon            724   3482
Jim Harbaugh          1076   2842
Rodney Peete           660   1686
Vinny Testaverde      1802   4727
Brett Favre              4   8219

1985 AFC Central       233   4027
Bernie Kosar             0   3365
Mark Malone            380   1268
Warren Moon            450   6373
Boomer Esiason         102   5103

1986 AFC Central       555   3705
Bernie Kosar           248   3117
Boomer Esiason         533   4672
Warren Moon            827   5996
Mark Malone            613   1035

1988 AFC East         1006   3573
Dan Marino            2494   5864
Ken O'Brien           1566   2036
Doug Flutie             71   2080
Jim Kelly              899   3880
Chris Chandler           0   4005

1983 AFC West         1401   3458
Jim Plunkett          2769    932
Dave Krieg             192   5119
Dan Fouts             3533   2071
Bill Kenney            512   1918
John Elway               0   7250

1986 AFC East          589   3401
Ken O'Brien            691   2911
Jim Kelly                0   4779
Dan Marino            1427   6931
Jack Trudeau             0   1644
Tony Eason             825    739

1988 AFC Central      1091   3310
Bubby Brister           72   2140
Warren Moon           1683   5140
Boomer Esiason        1442   3763
Bernie Kosar          1168   2197

1990 NFC Central       535   3246
Rich Gannon             21   4185
Jim Harbaugh           286   3632
Don Majkowski         1062    843
Vinny Testaverde      1111   5418
Rodney Peete           195   2151

1987 AFC Central      1034   3227
Boomer Esiason        1002   4203
Mark Malone           1038    610
Bernie Kosar           779   2586
Warren Moon           1315   5508

1984 AFC West         1018   3207
Dan Fouts             3873   1731
John Elway             259   6991
Todd Blackledge         34    847
Marc Wilson            490   1591
Dave Krieg             435   4876

1991 NFC Central       690   3168
Rich Gannon            370   3836
Vinny Testaverde      1476   5053
Mike Tomczak           915   1422
Jim Harbaugh           598   3320
Erik Kramer             92   2207

1995 AFC Central      1161   3159
Mark Brunell            27   4567
Chris Chandler        1242   2763
Neil O'Donnell        1455   1774
Jeff Blake             315   2926
Vinny Testaverde      2766   3763

1993 AFC East         1507   3096
Scott Mitchell           8   2338
Drew Bledsoe             0   6717
Boomer Esiason        3378   1827
Jim Kelly             3024   1755
Jeff George           1125   2842

1988 NFC West         1093   2989
Chris Miller            92   2800
Joe Montana           3276   2115
Bobby Hebert           554   2567
Jim Everett            449   4474

1987 AFC East         1412   2983
Dan Marino            2050   6308
Steve Grogan          2939    654
Jim Kelly              480   4299
Jack Trudeau           417   1227
Ken O'Brien           1173   2429

1989 AFC Central      1419   2982
Boomer Esiason        1830   3375
Warren Moon           1977   4846
Bernie Kosar          1427   1938
Bubby Brister          442   1770

No reason to stop with quarterbacks. The 1998 AFC Central housed a lot of future running back production. Priest Holmes and Fred Taylor were in their first seasons, Corey Dillon was in his second, Eddie George his third, and even the old man Jerome Bettis wasn't yet halfway done with his career. The numbers shown here are rushing yards.

1998 AFC Central      2017   8562
Corey Dillon          1129  10112
Fred Taylor              0   9513
Jerome Bettis         6187   7475
Priest Holmes            0   8035
Eddie George          2767   7674

1996 AFC Central      1250   7041
Garrison Hearst       1239   6463
Eddie George             0  10441
Jerome Bettis         3091  10571
James Stewart          525   5316
Bam Morris            1395   2414

1995 AFC East         2249   6917
Bernie Parmalee        922   1257
Curtis Martin            0  14101
Adrian Murrell         317   4882
Marshall Faulk        1282  10997
Thurman Thomas        8724   3350

1993 NFC West          363   6909
Erric Pegram           438   2960
Derek Brown              0   1383
Jerome Bettis            0  13662
Ricky Watters         1013   9630

1995 NFC East         3559   6883
Ricky Watters         2840   7803
Garrison Hearst        169   7533
Terry Allen           2795   5819
Rodney Hampton        4807   2090
Emmitt Smith          7183  11172

1979 NFC East         1931   6629
Billy Taylor           250   1394
John Riggins          5669   5683
Tony Dorsett          2332  10407
Wilbert Montgomery    1403   5386
Ottis Anderson           0  10273

1997 AFC Central      2251   6623
Eddie George          1368   9073
Jerome Bettis         4522   9140
Bam Morris            2132   1677
Corey Dillon             0  11241
Natrone Means         3232   1983

1997 AFC East         1813   6562
Curtis Martin         2639  11462
Adrian Murrell        2361   2838
Karim Abdul-Jabbar    1116   2297
Antowain Smith           0   6881
Marshall Faulk        2947   9332

As is often the case, the wide receiver list degenerates into a whole page of Jerry Rice-fueled groups. I'll show the top division, and then some others of interest. Lots of Rice groups have been deleted in between.

1986 NFC West         1419  10676
Eric Martin            522   7639
Charlie Brown         2527   1021
Henry Ellard          1701  12076
Jerry Rice             927  21968

1998 AFC East         1802   7761
Eric Moulds            573   9075
Terry Glenn           1563   7260
Marvin Harrison       1702  11995
Keyshawn Johnson      1807   8764
O.J. McDuffie         3365   1709

2000 NFC West         2623   6833
Joe Horn               879   7622
Torry Holt             788   9887
Muhsin Muhammad       2918   6446
Terrell Owens         3307   8408
Shawn Jefferson       5222   1801

1997 AFC West         4469   6561
Tony Martin           5002   4063
Joey Galloway         2026   7532
Rod Smith              389  11000
Tim Brown             7180   7754
Andre Rison           7747   2458

1991 AFC East         4290   6154
Andre Reed            5353   7845
Mark Duper            7022   1847
Irving Fryar          3921   8864
Rob Moore              692   8676
Billy Brooks          4462   3539

1995 AFC West         3303   6138
Joey Galloway            0   9558
Anthony Miller        6689   2459
Willie Davis          2487   2016
Tony Martin           2607   6458
Tim Brown             4734  10200

1978 AFC West         2337   6087
John Jefferson           0   5714
Henry Marshall         888   5657
Steve Largent         1348  11741
Haven Moses           5484   2607
Cliff Branch          3967   4718

Finally, the tight ends:

1997 AFC West         1160   4241
Freddie Jones            0   4232
Carlester Crumpler     531    448
Shannon Sharpe        4884   5176
Rickey Dudley          386   2638
Tony Gonzalez            0   8710

1998 AFC West         1708   3995
Tony Gonzalez          368   8342
Shannon Sharpe        5991   4069
Rickey Dudley         1173   1851
Freddie Jones          505   3727
Christian Fauria       505   1985

1978 AFC Central       158   3308
Ozzie Newsome            0   7980
Don Bass                 0   1580
Mike Barber             94   2694
Randy Grossman         537    977

1979 AFC Central       590   3217
Bennie Cunningham      717   2162
Mike Barber            607   2181
Ozzie Newsome          589   7391
Don Bass               447   1133

1999 AFC West          979   3122
Rickey Dudley         1722   1302
Christian Fauria       882   1608
Freddie Jones         1107   3125
Tony Gonzalez          989   7721
Byron Chamberlain      193   1855

1980 AFC Central      1018   2907
Ozzie Newsome         1370   6610
Randy Grossman        1202    312
Mike Barber            984   1804
Dan Ross               516   2903

1982 AFC West          745   2897
James Wright            48    618
Kellen Winslow        2620   4121
Al Dixon               943    305
Pete Metzelaars          0   3686
Todd Christensen       115   5757

1981 AFC Central      2114   2760
Dan Ross              1240   2179
Dave Casper           3791   1425
Bennie Cunningham     1461   1418
Ozzie Newsome         1964   6016

As careful readers of this blog will know, I am not a fan of Bill Parcells. I'll spare you the reasons for that, mainly because I don't even remember them very well myself, and make this a Friday Discussion Question instead of a Friday Rant.

Because of my anti-Parcells agenda, I have been vigorously rooting against the Cowboys for the last several years. Now that he's moved on, what's a hater to do?

If the Cowboys struggle, does that mean Parcells' influence was the only thing keeping them near .500? Or does it mean he did a crummy job of building the foundation of the team? If the Cowboys go to the Super Bowl, will that be because Parcells built a solid nucleus that Wade Phillips just had to not ruin, or because Phillips finally was able to get this talented team to play to its potentential, something it had not been able to do under the stifling Parcells regime?

In general, what --- if anything --- do we learn about a coach the year after he leaves? I'm sure it varies from case to case, but what kinds of things does it depend upon? What specific events, if any, could happen in Dallas this year that would change your opinion of the job Parcells did there?

In this thead at the footballguys message board, a poster named "Lash" points out that, because they play in a division where the other starting quarterbacks could be JaMarcus Russell, Brodie Croyle, and Jay Cutler, the San Diego Chargers figure to get lots of games this year against inexperienced quarterbacks.

That, combined with the fact that the Chargers themselves don't exactly have a ten-year vet at the helm, prompted me to see if any division has ever had such a collective lack of experience at the quarterback position. It turns out that, even if Josh McCown and/or Damon Huard start in place of Russell and/or Croyle, this will be among the greenest quarterback divisions of the last 30 years.

I looked at all divisions since 1978, defined their quarterback to the one who threw the most passes for them during that season, and then measured that quarterback's experience by counting the number of NFL passes he had thrown prior to that season. Then I averaged those figures for all teams in the division.

That sounds complicated, but it's not. Let me illustrate with the 2007 AFC West:

   QB            Career passing attempts
========================================
Rivers                   490
Cutler                   137
Croyle                     7
Russell                    0

The average is 159, which would be by far the lowest since 1978. Even if we slot Huard and McCown in there in place of Croyle and Russell, we'd get 505, which is still very low.

1985 AFC Central       233
Bernie Kosar             0
Mark Malone            380
Warren Moon            450
Boomer Esiason         102

1979 NFC Central       421
Doug Williams          194
Jeff Komlo               0
Tommy Kramer            73
Mike Phipps           1405
David Whitehurst       433

1988 NFC Central       495
Jim McMahon           1321
Wade Wilson            703
Rusty Hilger           157
Vinny Testaverde       165
Don Majkowski          127

1980 NFC Central       525
Doug Williams          591
Tommy Kramer           639
Lynn Dickey            876
Gary Danielson         451
Vince Evans             66

1989 NFC Central       533
Mike Tomczak           505
Bob Gagliano            30
Wade Wilson           1035
Don Majkowski          463
Vinny Testaverde       631

1990 NFC Central       535
Rich Gannon             21
Jim Harbaugh           286
Don Majkowski         1062
Vinny Testaverde      1111
Rodney Peete           195

1986 AFC Central       555
Bernie Kosar           248
Boomer Esiason         533
Warren Moon            827
Mark Malone            613

1986 AFC East          589
Ken O'Brien            691
Jim Kelly                0
Dan Marino            1427
Jack Trudeau             0
Tony Eason             825

2004 NFC West          594
Tim Rattay             163
Matt Hasselbeck       1282
Marc Bulger            746
Josh McCown            183

Here are the most experienced divisions full of quarterbacks:

2006 NFC North        3610
Brad Johnson          3797
Jon Kitna             2837
Rex Grossman           195
Brett Favre           7610

1995 AFC East         3473
Dan Marino            6049
Drew Bledsoe          1120
Boomer Esiason        4291
Jim Harbaugh          1961
Jim Kelly             3942

2005 NFC East         3178
Donovan McNabb        2586
Mark Brunell          3880
Drew Bledsoe          6049
Eli Manning            197

1997 AFC West         3107
Craig Whelihan           0
Warren Moon           6000
John Elway            6392
Jeff George           2712
Elvis Grbac            430

1996 AFC East         3085
Jim Harbaugh          2275
Jim Kelly             4400
Frank Reich            461
Drew Bledsoe          1756
Dan Marino            6531

2005 AFC West         3056
Kerry Collins         4517
Drew Brees            1309
Trent Green           2822
Jake Plummer          3577

1998 AFC West         3036
Rich Gannon           1404
John Elway            6894
Donald Hollas          115
Craig Whelihan         237
Warren Moon           6528

2005 NFC North        2995
Joey Harrington       1472
Kyle Orton               0
Brad Johnson          3503
Brett Favre           7003

1994 AFC East         2993
Jim Harbaugh          1759
Dan Marino            5434
Drew Bledsoe           429
Jim Kelly             3494
Boomer Esiason        3851

1998 AFC East         2974
Doug Flutie            341
Drew Bledsoe          2901
Peyton Manning           0
Vinny Testaverde      4177
Dan Marino            7452

1981 AFC Central      2868
Ken Anderson          3060
Ken Stabler           2938
Terry Bradshaw        3283
Brian Sipe            2191

1997 AFC East         2830
Drew Bledsoe          2379
Neil O'Donnell        2059
Dan Marino            6904
Todd Collins           128
Jim Harbaugh          2680

I got an email recently from a guy named Matt who runs a college sports blog called Statistically Speaking. In his email he referred to a recent post of his where he unveiled an interesting modification of the Pythagorean Theorem for football. First let me tell you just a bit about what the theorem is.

In the early 80s, or possibly even before that, Bill James noted that baseball teams' true strengths could generally be measured more accurately by looking at runs scored and runs allowed than by looking at wins and losses. To be more precise, he found that one can predict future win/loss records more accurately using only past runs scored and runs allowed than using only past wins and losses. To put it another way, if a team had a record of 82-80, but their runs scored and allowed totals were more in line with those of a 76-86 team, then that team should be treated as a 76-86 team for the purposes of predicting next year's record.

So what record "should" a team with RS runs scored and RA runs allowed have had? James came up with the formula:

                       RS^2
Expected record =~  -----------
                    RS^2 + RA^2

Because it has some superficial similarities to the Pythagorean Theorem about right triangles that you learned at some point in your youth, it came to be known by the same name. As it turns out, though, you can replace the 2s in the exponents with 1.82s and get slightly better predictions. For football, people have found that an exponent of 2.37 seems to work best. So the Pythagorean Theorem for football looks like this:

                          PF^2.37
Expected record =~  -----------------
                    PF^2.37 + PA^2.37

Here is each team's Pythagorean record for 2006.

Tm   record  PythagRecord
=========================
sdg   14- 2   12.1- 3.9
bal   13- 3   12.7- 3.3
chi   13- 3   12.4- 3.6
nwe   12- 4   12.2- 3.8
ind   12- 4    9.6- 6.4
phi   10- 6    9.8- 6.2
nor   10- 6   10.3- 5.7
nyj   10- 6    8.7- 7.3
sea    9- 7    7.8- 8.2
kan    9- 7    8.5- 7.5
den    9- 7    8.4- 7.6
dal    9- 7    9.8- 6.2
ten    8- 8    6.0-10.0
jax    8- 8   10.8- 5.2
stl    8- 8    7.6- 8.4
nyg    8- 8    7.8- 8.2
car    8- 8    6.9- 9.1
gnb    8- 8    6.2- 9.8
pit    8- 8    9.1- 6.9
cin    8- 8    9.1- 6.9
atl    7- 9    6.9- 9.1
buf    7- 9    7.7- 8.3
sfo    7- 9    5.1-10.9
min    6-10    6.6- 9.4
mia    6-10    7.2- 8.8
hou    6-10    5.1-10.9
was    5-11    6.1- 9.9
ari    5-11    6.0-10.0
cle    4-12    4.4-11.6
tam    4-12    3.6-12.4
det    3-13    5.6-10.4
oak    2-14    2.7-13.3

Here is a quick demonstration of the method's ability to predict the future on the group level. I looked at all teams since 1978 with a record of exactly 10-6. Then I divided them into three groups: [1] those with 9.5 or fewer Pythagorean wins (these were the teams we might say were lucky to be 10-6), [2] those with between 9.5 and 10.5 Pythagorean wins (these teams really were morally 10-6 teams), and [3] those with 10.5 or more Pythagorean wins (these, we would speculate, were actually stronger than 10-6 teams). Here's how they did the next year:

                          Average next year wins
=================================================
9.5 or fewer PWins                 7.9
9.5--10.5 PWins                    9.3
10.5 or more PWins                 9.9

Perhaps a more rigorous proof of the method's power is this regression of Year N+1 wins on the two variables: Year N Wins, and Year N Pythagorean wins:

Predicted Year N+1 wins =~ 4.07 + .12*(Year N Wins) + .38*(Year N Pythag wins)

The coefficient on Pythagorean wins is much larger than that on actual wins, which says that Pythagorean wins are more closely associated with Year N+1 wins than are actual wins. Further, regression fans will want to know that the coefficient on actual wins is only barely significant if at all (p=.07), while the coefficient on Pythagorean wins is highly significant (p=.0000009).

So, we finally get to Matt's new method. I'll quote from his blog post:

blowouts, especially extreme blowouts can artificially inflate or deflate a team's Pythagorean record depending on whether or not they received or doled out the beating. The solution? Compute the Pythagorean winning percentage on a game by game basis, add up the totals, and divide by games played. This way each game is counted the same and the effect of blowouts is lessened.

Here are the 2006 records, Pythagorean records, and adjusted (by Matt) Pythagorean records:

Tm   record     Pythag    New Pythag
====================================
sdg   14- 2   12.1- 3.9   11.5- 4.5
bal   13- 3   12.7- 3.3   11.1- 4.9
chi   13- 3   12.4- 3.6   11.1- 4.9
nwe   12- 4   12.2- 3.8   10.5- 5.5
ind   12- 4    9.6- 6.4    9.3- 6.7
phi   10- 6    9.8- 6.2    9.5- 6.5
nor   10- 6   10.3- 5.7    9.8- 6.2
nyj   10- 6    8.7- 7.3    8.7- 7.3
sea    9- 7    7.8- 8.2    8.4- 7.6
kan    9- 7    8.5- 7.5    8.0- 8.0
den    9- 7    8.4- 7.6    9.3- 6.7
dal    9- 7    9.8- 6.2    9.3- 6.7
ten    8- 8    6.0-10.0    7.0- 9.0
jax    8- 8   10.8- 5.2    9.9- 6.1
stl    8- 8    7.6- 8.4    7.7- 8.3
nyg    8- 8    7.8- 8.2    8.4- 7.6
car    8- 8    6.9- 9.1    7.9- 8.1
gnb    8- 8    6.2- 9.8    7.3- 8.7
pit    8- 8    9.1- 6.9    8.1- 7.9
cin    8- 8    9.1- 6.9    9.3- 6.7
atl    7- 9    6.9- 9.1    7.2- 8.8
buf    7- 9    7.7- 8.3    8.2- 7.8
sfo    7- 9    5.1-10.9    7.0- 9.0
min    6-10    6.6- 9.4    6.6- 9.4
mia    6-10    7.2- 8.8    7.2- 8.8
hou    6-10    5.1-10.9    6.5- 9.5
was    5-11    6.1- 9.9    6.5- 9.5
ari    5-11    6.0-10.0    6.2- 9.8
cle    4-12    4.4-11.6    5.0-11.0
tam    4-12    3.6-12.4    4.4-11.6
det    3-13    5.6-10.4    5.5-10.5
oak    2-14    2.7-13.3    3.8-12.2

Now, the question is: if you know a team's wins, a team's Pythagorean wins, and a team's adjusted Pythagorean wins, what is the relative importance of each of those in predicting the team's Year N+1 wins? Here is what regression says:

Year N+1 wins =~ 3.93 + .11*(Year N wins) + .34*(Year N Pythag wins) + .06*(Year N AdjPythag wins)

The Pythag Wins coefficient is significant and the other two are not. All three inputs are, of course, very highly correlated, which can sometimes cause problems in regressions. What I don't know about the fine points of regression analysis could fill a warehouse, but I think we've got enough data here to conclude that the significance of the regular Pythagorean wins coefficient and the lack of significance for the other two means that Pythagorean wins is generally the better predictor in cases where there is some disagreement among the three.

One more try: a regression of Year N+1 wins on Year N wins and Pythagorean wins has an R^2 of .203. A regression of Year N+1 wins on Year N wins and Year N adjusted Pythagorean wins has an R^2 of .199.

So it appears to me that, for NFL games, the good old fashioned Pythagorean Theorem is no worse, and possibly a tiny, tiny, tiny, tiny bit better, than Matt's adjusted version. However, based on some preliminary investigations, Matt found the opposite in college football.

Is one of us wrong? I don't think so. It seems believable that deflating blowouts a bit would create a better gauge of team strength for college football teams and a (very slightly) worse one for NFL squads. Virginia Tech is so much better than Duke that they can essentially beat them by whatever score they want. Whether they decide to beat them 31-3 or 61-3 doesn't tell us anything more than what we already knew: the Hokies are much, much better. But, as bad as the Raiders may be, there are no Dukes in the NFL, and just about any team has a credible shot at beating any other on a Given Sunday. If you hang a severe blowout on an NFL team, it apparently says something about the relative strength of your teams.

At long last, my other site (footballguys.com) has a blog. We've got a large staff of writers with diverse interests, and we plan to crank out several posts a day during the preseason and on into the season. So check it out if you're looking for some fantasy-related reading material.

I'll still be posting here as usual, and I'll cross-post some general interest articles, but most of my fantasy football writing will be over there from now on.

Have I mentioned that I enjoy listening to the footballguys.com podcast: The Audible? Well I do, and you probably would too, so give it a try if you're a podcast kind of a person. It's free all year round. As I've mentioned before, they are currently going around the league and interviewing a beat writer from each NFL team. Their chat with Jags' writer Vito Stellino of the Florida Times-Union contained a number of interesting nuggets, among them was Vito's assertion that Maurice Jones-Drew benefited a lot last season from nickel defenses that were expecting the pass.

I decided to check it out, and ultimately I decided that Stellino was both right and wrong. Compared to his backfield-mate Fred Taylor, Jones-Drew did indeed see more carries in situations where passing is to be expected. Compared to the league as a whole, however, Jones-Drew's carries were not unduly tilted toward those situations.

For a simple look at things, note that Fred Taylor had only five (5) third-down carries all season, while Jones-Drew had 30. So it's not unreasonable to get the impression that Jones-Drew was running against a lot of thin front lines. However, it's tough to make a case that his numbers were a product of that fact. Consider that six of his nine longest runs (including the longest three) came on first-and-ten. Here is a log of all his 10+ yard rushes. The clock column runs from 0 (the opening kickoff) to 60 (the final gun). The score indicates the relative score, from the Jags' standpoint, at the time of the run.

d-d  clock score length
=======================
1-10  16.8   -3   74
1-10  28.5   +4   48
1-10  44.0  +37   40
2- 7  52.3  +10   32
1-10   5.1   +0   26
2- 5  18.8  -10   24
1-10  31.1  +21   22
2-25  11.5   +7   18
1-10   2.6   +0   18
2- 3  22.2   -3   17
2- 7  27.5  -10   17
3- 4  24.4   +3   17
1-10  22.9   -3   17
2-10  36.0   +0   15
1-10  19.4  -10   14
3- 3  58.0   -7   13
2-11  47.8   +7   13
2-16  56.1  -10   13
1-20  53.0  +10   12
1-10  35.1  +24   12
2- 1  39.1   +3   12
3- 7  12.2   +7   12
3-20  24.8   -4   12
1-10  20.5   +0   12
2-12  11.9   -7   12
2- 7  44.7   +7   11
1-10  28.1   +4   11
1-10  17.5   +0   10

A quick eyeballing doesn't reveal a preponderance obvious passing situations there.

But the p-f-r blog isn't about eyeballing for preponderances. I decided to quantify the "average situation" of all of Jones-Drew's rushing attempts and compare it to that of other running backs. Here's the plan.

Step 1: build a model that eats up the relevant situational variables and spits out the probability of a running play occurring. That wasn't pretty, but ultimately I got it built. I'll put the details at the end of the post for those who are interested. For those who just want the gist, here are a few examples:

• 2nd-and-6, down by 10, 5 minutes left in the second quarter: 40% probability of a run / 60% probability of a pass
• 2nd-and-6, down by 6, middle of the first quarter: 46% probability of a run / 54% probability of a pass
• 2nd-and-6, up by 6, late in the second quarter: 52% probability of a run / 48% probability of a pass
• 2nd-and-6, down by 39, start of the fourth quarter: 10% probability of a run / 90% probability of a pass
• 2nd-and-6, up by 14, start of the fourth quarter: 65% probability of a run / 35% probability of a pass

The model isn't perfect, but I think it's good enough for our purposes here.

Step 2: I looked at each rush of Maurice Jones-Drew's (and everybody else's) season and computed the average rush probability of those situations. So a runner who was running more often in passing situations, and/or in more extreme passing situations, would have a low value. While runners who were often running while ahead and in classic running situations would have a high value. Here is the list of all players with at least 100 rushes:

Cedric Houston         0.560
Marion Barber III      0.553
Cedric Benson          0.545
Deuce McAllister       0.535
Dominic Rhodes         0.518
Jamal Lewis            0.514
Joseph Addai           0.512
Corey Dillon           0.511
Laurence Maroney       0.506
Thomas Jones           0.504
LaDainian Tomlinson    0.497
Julius Jones           0.486
Kevan Barlow           0.484
Fred Taylor            0.484
Steven Jackson         0.483
Willie Parker          0.483
DeAngelo Williams      0.482
Rudi Johnson           0.482
DeShaun Foster         0.480
Edgerrin James         0.479
Shaun Alexander        0.477
Willis McGahee         0.477
Maurice Morris         0.474
Mike Bell              0.472
Ronnie Brown           0.471
Larry Johnson          0.468
Maurice Jones-Drew     0.467
Clinton Portis         0.466
Kevin Jones            0.464
Wali Lundy             0.462
Reggie Bush            0.461
Chester Taylor         0.460
Tiki Barber            0.460
Ahman Green            0.457
Brian Westbrook        0.454
Anthony Thomas         0.451
Justin Fargas          0.450
Frank Gore             0.447
Tatum Bell             0.445
Ron Dayne              0.444
Warrick Dunn           0.442
Travis Henry           0.433
Reuben Droughns        0.428
Cadillac Williams      0.425
Ladell Betts           0.424
Leon Washington        0.418
LaMont Jordan          0.408
Michael Vick           0.407

As you can see, Jones-Drew is below Taylor, but completely unremarkable compared to the rest of the league's backs. And, as we saw above, Jones-Drew's big runs did not come in passing situations anyway. Here is a list of the "average situation" of all running backs on their runs of ten or more yards only (minimum 10 such runs):

Michael Turner         0.532
Vernand Morency        0.511
Correll Buckhalter     0.509
Dominic Rhodes         0.506
DeAngelo Williams      0.503
Jamal Lewis            0.502
Laurence Maroney       0.493
Willie Parker          0.493
Marion Barber III      0.491
Ronnie Brown           0.488
Deuce McAllister       0.484
Willis McGahee         0.483
Cedric Benson          0.483
Rudi Johnson           0.482
Joseph Addai           0.478
Leon Washington        0.477
Fred Taylor            0.476
Julius Jones           0.474
Maurice Morris         0.471
Frank Gore             0.471
Mike Bell              0.471
Edgerrin James         0.468
LaDainian Tomlinson    0.467
DeShaun Foster         0.465
Jerious Norwood        0.458
Kevin Jones            0.458
Thomas Jones           0.456
Wali Lundy             0.455
Travis Henry           0.453
Steven Jackson         0.450
Corey Dillon           0.447
Chester Taylor         0.441
Brian Westbrook        0.437
Larry Johnson          0.437
Maurice Jones-Drew     0.431
Tiki Barber            0.430
Reggie Bush            0.424
Warrick Dunn           0.422
Ladell Betts           0.417
Brandon Jacobs         0.414
Justin Fargas          0.413
LaMont Jordan          0.392
Michael Vick           0.391
Shaun Alexander        0.386
Clinton Portis         0.382
Ahman Green            0.381
Ron Dayne              0.375
Reuben Droughns        0.374
Cadillac Williams      0.374
Tatum Bell             0.359
Donovan McNabb         0.288
Vince Young            0.246

Maurice is a little on the low side, but essentially in the middle of the pack.

In conclusion:

• I see no evidence that Maurice Jones-Drew faced, all things considered, a particularly easy set of situations to run in.
• I see no evidence that he took undue advantage of the easy situations he did see.

So, while I do think Jones-Drew's 5.7 yards per rush average last season was partly a result of luck (no one is truly a 5.7 yards per rusher), I do not think it was a result of situation.

Details of the model: as you might guess, the model is based on a logistic regression of leaguewide play-by-play data. Here it is:

Run probability =~ 1 - 1/(1 + exp(-.1201 + .001135*(ScoreDifferential)*(ClockTime) + X))

Where ClockTime = 0 at the start of the game and 60 at the end of the game, and X takes on the following values:

1st,10:            0
1st,11+:      -.3105
1st,9-:        .7687
2nd,0-3:       .7675
2nd,4-6:       .0061
2nd,7-9:      -.6187
2nd,10+:      -.5225
3rd/4th,0-1:  1.0670
3rd/4th,2-3:  -.8121
3rd/4th,4-7: -1.9350
3rd/4th,8+:  -1.7210

Note that the more positive the number is, the more likely teams are to run in that situation.

I don't have offensive linemen or punters in my database, and I'm missing a few birthdates here and there, but those caveats aside, here is the full list of players who played in the NFL in 2006 and were born on a Friday the 13th:

• Brad Johnson
• Jason Craft
• Landon Johnson
• L.P. LaDouceur

In addition, the following players turned 13 years old on a Friday the 13th:

• Michael Clayton
• Jerramy Stevens
• David Martin
• Ahmard Hall
• Sam Adams
• Shaun Phillips

Now consider:

• Michael Clayton sprained his knee in week 13 of last season and never again returned.
• Brad Johnson's week 13 game included no TD passes and four interceptions. Although he did play in the following weeks, this game ultimately set in motion the chain of events that led to his benching. In fact, Johnson's TD/INT ratio is 11/12 in week 13 and 153/105 in all other weeks.
• In his six years in the league, David Martin has only played in week 13 once. Only 17 of his 766 career receiving yards have come in week 13.
• Sam Adams is incredibly fat.

But here is the clincher:

In week 13, Ahmard Hall had 2 rushes for 13 yards. That week, his Titans beat the eventual Super Bowl Champion Colts, but only just barely.

This post will continue with looking at workloads and evidence of overuse, by examining the end of the season games, and playoffs, and what effect it has on injury rates early the next season.

The below chart contains all running backs who played in all team games between weeks 12-17, with a couple of caveats. First, if a back missed only week 17, and he played in the playoffs, he was included. Also, between 1995-2001, there were a few backs that had bye weeks at the end of the year. Thus, we are not dealing with backs who had the same number of games, some could have five, some could have six. Thus, the players are divided by attempts/game, and not raw attempts, like yesterday.

This chart also does not separate out playoff participants from non-playoff participants (I'll discuss the playoff games shortly). Here are the injury rates for the early part of next season, sorted by average attempts per game over the last 6 weeks of the regular season. "GP" is the average number of games played the next season by players in that group. "SEI" represents "season ending injury" and is for all backs who had a season ending injury within the first 6 games of the following season. "INJ6" represents the running backs who missed at least 1 game due to injury in the first 6 games of the next season, but returned to play following the injury. If you add "SEI" and "INJ6", you will have the total percentage of players who missed at least one game out of the first six the following season. "PL15" represents all backs, among those who made it through the first 6 games of the next year healthy, who went on to play in at least 15 regular season games.

Att/G          No        GP       SEI     INJ6     PL15
=======================================================
25.0+          14       11.0     0.286   0.214    0.857
23.0-24.9      23       13.3     0.087   0.217    0.733
21.0-22.9      41       13.7     0.098   0.073    0.794
19.0-20.9      41       14.3     0.024   0.122    0.771
17.0-18.9      34       13.3     0.059   0.294    0.682
15.0-16.9      37       13.4     0.027   0.270    0.593
=======================================================

The pattern is similar to the early season pattern from the previous post. The groups between 19.0 and 22.9 attempts per game represent the peak in terms of average games played the next season. For the lower carry groups, the season ending injury rate is not high, but the players in these groups tend to miss some games early, and those who do not miss some games later at a higher frequency. The higher carry groups have increased season ending injury rates in the short term, but if they survive the early part of the next season without injury, they tend to remain healthy at a high rate.

Here are the players included in the above study, but sorted by games played in that year's postseason, rather than rushing averages:

Playoffs   No      GP     SEI     INJ6    PL15
===============================================
0         104     13.5   0.077   0.192   0.750
1          43     13.8   0.047   0.163   0.758
2          25     13.4   0.040   0.200   0.632
3+         18     12.1   0.167   0.222   0.667
===============================================

When the players are sorted by playoff games played after the season, there is generally no effect on injury rates, with a couple of exceptions. As we have seen, higher workloads increase the risk of serious injury already. Extending the season in the playoffs is like a spark, and high workloads are the gas. Also, if a team is using a running back heavily in the final weeks of the regular season, they do not typically reduce that use in the playoffs, so we are generally seeing the overuse period extended.

Three running backs averaged 25.0 or more rush attempts per game over the final six weeks of the regular season, and then played in three or more playoff games. Those three are Terrell Davis (1998), Jamal Anderson (1998), and Jamal Lewis (2000). I do not need to spend much time discussing their well-known injuries. In short, all three blew out knees early the next year: Lewis before the season began, Anderson in game two, and Davis in game four. If we add in the 23.0-24.9 group, four more running backs played in at least three playoff games. Those backs were Emmitt Smith (1995), Dorsey Levens (1997), Corey Dillon (2004) and Shaun Alexander (2005). Emmitt played in 15 games the next season; the other three all suffered injuries early the next season. Both Dorsey Levens and Shaun Alexander had serious injuries early, had consecutive games missed, but did return later in the season. For Levens, it was nine games missed in a row with a knee injury. Alexander missed six consecutive games last season with a foot injury.

The other exception where extending the playoffs increased injury rates involved recently injured backs. I found thirty backs who played in at least 3 playoff games (and averaged at least 10 carries a game) in one post season since 1995. Of those thirty, six had missed at least one game in weeks 12-16 of the regular season, before returning in the playoffs. Of those six, five of them missed at least one game within the first six games of the next season, and none of them played in more than 13 games the next year. The six were Lamont Warren (1995), Bam Morris (1995), DeShaun Foster and Stephen Davis (2003), Antowain Smith (2003), and Brian Westbrook (2004). Of those, two suffered season ending injuries early--the Panthers' Davis and Foster. Davis played two games before knee injury ended his season. Foster played in four before a knee injury--having a 32 attempt game in week 2 probably did not help either.

In my opinion, it is not the raw number of carries that matters. I believe the key is the number of higher stress games a back endures over a period of time. I will refer to these as Increased Risk Games (IRG). For now, my educated guess in reviewing the data is that the cutoff to qualify for an IRG is around 25 rush attempts for an average, healthy starting running back. More research is needed in this area, as the number may be lower or slightly higher depending on other factors, such as age, recent injury history, or stamina/conditioning of the specific back.

The pro football reference database has regular season game by game data back to 1995, but the playoff database goes back to the start of the Super Bowl era. Going back to 1978, there were 74 player-seasons where a back had at least one IRG in a postseason. Ten of those 74 backs (13.5%) played in six or fewer games the next season. Four of them were already discussed--Lewis, Anderson, Terrell Davis, and Stephen Davis. Going back in time, the other six were Edgar Bennett (1996)-0 games, Greg Bell (1989)-6 games, Ickey Woods (1988)-2 games, Dan Doornink (1984)-6 games, Curt Warner (1983)-1 game, and Wendell Tyler (1979)-4 games. Rob Carpenter (1981) also played in less than 6 games the next season, but the next season was the strike-shortened 1982 season, and I could not find information on the reason for his missed games to know whether to include him.

Here are some other recent cases that were not previously discussed, or not included in the previous studies, but which also show the potential role IRG plays in increasing risk of injuries, beyond what simply knowing raw rushing attempt totals might tell us. I could have selected numerous others; I selected these to get a mix of different type players or players with different reputations.

• Samkon Gado (2005): his 8 games/143 attempts might not seem excessive, but he had 4 IRG in the 6 games before his season-ending knee injury.
• Deuce McAllister (2005): 82 carries through 4 games prior to ACL injury does not seem excessive, but two of the four games were IRG.
• Clinton Portis (2003): Portis had 4 IRG in last 6 games, including 38 attempts in final game, before missing the final 2 games of the regular season with chest injury.
• Lee Suggs (2003 and 2004): After being injured and little used all year, Suggs had 26 carries in week 17 of 2003. He missed first 3 games next season. He had back to back IRG at the end of 2004, including 38 attempts in week 16. Played in only two of first ten team games next season, and has had only 14 attempts last two seasons.
• Duce Staley (2000): Staley's 79 carries through 5 games, and 326 carries in 16 games the previous year might not suggest overuse, if you were going by raw attempts. He closed 1999 with 2 of 4 IRG, and then had 26 attempts in game 1 of 2000. He went downhill quickly thereafter, and was done by week 5.

If you want to know how the regression Doug ran last year, and the research here can be reconciled, I believe, in a figurative sense, it is because of Samkon Gado and Lee Suggs. I am skeptical that prior historical workload has anything to do with injury risk. What matters is recent workload. In other words, Curtis Martin may have broken down in 2005 beyond what could be expected on age, due to workload over the second half of 2004, but I do not think what he did from 1995-2000 had any impact on whether he was going to break down or not. If Larry Johnson breaks down early in 2007, it will be because of what happened late in 2006, and have little to do with what happened late in 2005.

So, for Johnson, there is a glimmer of hope. If his workload is reduced next year, and he does not show signs of an injury early in the season, he is probably going to be okay (assuming no more IRG, an assumption not likely to be realized). I will close with my assessment of the top five injury risks early next season. The top one is probably not who you think it would be, in a season where the rush attempt record was broken.

1. Shaun Alexander-- Alexander came off missing six games due to the foot injury, and was immediately subjected to a workload greater than any of his career. Apparently, Holmgren thought he needed to be punished for missing games. Alexander had a whopping 220 carries over 8 games since week 12 last year, including 5 IRG. With his age, recent injury history, extremely high IRG rate, and rumors of the foot injury not being healed, Alexander has more red flags than the United Nations.
2. Larry Johnson-everyone knows about his record setting rushing totals last year. Unsurprisingly, then, he had 4 IRG in his last 7 games. His one saving grace is that he was not extended in the playoffs, and the Colts bottled him early and prevented him from adding to that IRG total.
3. Steven Jackson- through 13 games last year, Jackson was a model of how a top running back should be used to keep him healthy and productive. Jackson had zero IRG, as he consistently had games of between 18 and 22 attempts. Then, for whatever reason, he finished with 3 IRG in his last 3 games, including over 30 attempts in two of them, which places him squarely in the high risk category for developing an injury early in 2007. Linehan will have major regrets if Jackson's season is ruined due to work while playing out the string on a non-playoff year in 2006.
4. Rudi Johnson- Rudi closed with 3 out of 6 IRG. That alone might place him in the group. Couple that with his age beginning to get on the wrong side of prime running back age, his decline in ypc last year to 3.8, and the Bengals using a 2nd round pick on a running back, and I have my concerns.
5. Ladell Betts- Betts played backup most of the season, but closed with 3 out of 6 IRG, and 156 attempts in his last 6 games. He may not get enough attempts to make him as big a risk early, but he is still a risk to breakdown early.

Now, am I predicting any of these guys to definitely get hurt? No. When I say high risk, that means each of them have about a 15-25% chance of suffering a significant injury early in 2007, based on the historical data. But that is a significantly higher chance than other running backs, such as Joseph Addai or Reggie Bush, have of suffering a severe injury early in 2007.

Last year, Doug looked at the question of running back deterioration: age vs. mileage, here, here, and here. In that series of posts, the issue of age vs. career rushing attempts was examined.

I am going to approach the issue on a narrower level, by digging into the individual games database to see if there is evidence of a danger zone when it comes to rushing attempts, measured over a shorter period of time. In baseball, for example, pitch counts have become a common statistic, and Rany Jazerli and Keith Woolner of Baseball Prospectus developed a statistic known as Pitcher Abuse Points (PAP). Here is the question I will seek to answer: Is there a "rush count" in football that should be observed over the course of a game, or period of games, that should be observed to keep running backs healthy?

Two potential issues arise when looking for evidence of overuse. The first is whether high attempts increase the risk of injury serious enough to cause missed games. The second is whether high attempts reduce effectiveness/performance beyond what would be normally expected. And of course, these two issues can be related, as a severe injury could reduce a back's effectiveness/performance faster than normal, even after the back has "returned from injury."

For now, I am going to focus on the first issue (injury), by looking at games played. "Games played" is not a perfect measurement for injury, as players could also be benched, or be suspended. However, by limiting the players examined to those averaging over 15 carries a game over a period of time, we can be fairly certain that most games missed are due to injury, rather than coaching decisions.

Using the games database for the 1995-2006 regular seasons, I ran a sort of all running backs who (1) played in all 6 of their team's first 6 games; and (2) averaged 15.0 or more carries per game over that span (90 attempts or more total). I then sorted them by attempt totals, and looked at the average number of games played for the rest of the season.

Thus, if a player started off the season missing some games, or even worse, suffering a severe injury, they are not included in this particular study--they will be addressed later. By including only players who played in all games through the first six, the hope is that we are looking at players who are healthy. Or at least healthy enough to continue to play. There is nothing magical about my choice of a six game period. The main reason I chose it was because it was a small enough span to capture a good amount of players but also large enough to see the effect of workload over a group of games. It also left a nice, round, ten regular season games for each player, so the math on the average games played is easy. Here are the results:

# Attempts     # Players   Av GP, rest of year
==============================================
150 or more         9             7.9
138 to 149          24*           8.5
126 to 137          42**          9.2
114 to 125          46            9.2
102 to 113          54            8.5
 90 to 101          37            9.1

* Lamar Smith, 2002, excluded because placed on leave due to a DUI.  He missed last 5 games of season.
**William Green, 2003, excluded due to missing last 9 games after league suspension
=================================================

We see that the running backs peak, in terms of number of remaining games played, in the 114 to 137 carries range (19.0 to 22.9 attempts/game). The average number of games played dips on either side of this group. However, if we look at how the "138-149 attempts" and "150+ attempts" groups ("the high carry groups") are missing games compared to the lower attempt groups, a different pattern emerges.

Here are the percentage of players for each group who, due to injury, played in 50% or less of the remaining ten team games:

• 150 or more 2/9 22.2%
• 138 to 149 3/24 12.5%
• 126 to 137 2/42 4.8%
• 114 to 125 2/46 4.3%
• 102 to 113 5/54 9.3%
• 90 to 101 1/37 2.7%

The high carry groups show a significant increase in the percentage of players missing at least half of the remaining games. But still, we are talking about five players, and a couple of freak injuries could skew the data. So, I took a closer look at those 15 players who missed 50% or more of the remaining games. The players are presented in order of total attempts through the first six games, with attempts and remaining games missed (out of ten) listed. I have also discussed the nature of each injury, when it developed, and notable issues of immediate workload around the time injury, if they existed.

1. Ricky Williams, 2000, age 23, 155 att, 6 games-- Williams broke an ankle in game 10 vs. CAR, ending his season. His workload was even higher right before the injury than it was through the first six. He averaged an astounding 29 attempts per game in the 5 games before the injury. Ricky had also missed 4 games in 1999 (rookie year).
2. Edgerrin James, 2001, age 23, 151 att, 10 games-- James tore his knee on a stretch play to the outside, on his 27th carry in a week 7 game in KC, after games of 26 and 30 attempts the previous two weeks, following a week 4 bye.
3. Priest Holmes, 2004, age 31, 148 att, 8 games-- Priest injured his knee in week 9 at TB, missed rest of season. Had 32 carries previous week, which was 3rd game out of previous 5 with more than 30 attempts.
4. Robert Smith, 1996, age 24, 140 att, 8 games-- Smith had missed 7 games previous year. Had 3 games of 26+ attempts in the 6 games immediately prior to injury. Tore knee in game 8 vs. CHI on the 4th carry.
5. Natrone Means, 1998, age 26, 139 att, 6 games-- Means (apparently worn out from playing with Ryan Leaf) suffered a season-ending foot injury early in week 11 vs. BAL. Had 27+ attempts in 3 of previous 5 games, including a game with 37 attempts. Had missed 11 games over previous 3 seasons to injury. Played in 7 more career games.
6. Natrone Means, 1995, age 23, 133 att, 6 games-- Means suffered a groin injury on the 2nd carry of week 10, coming off a bye week. Means had only 1 game with less than 22 carries since week 1, including 26 the last game before the injury. Came back in week 16, but re-aggravated injury after 3 carries, and missed the final week as well.
7. Duce Staley, 2004, age 29, 126 att, 6 games-- Staley had 25 carries in final game before suffering hamstring injury, his 2nd 25 carry game in 4 games. Missed 4 straight games, came back to play 3 of last 5, but only had 41 total attempts. He played in 6 more games after 2004 season.
8. Chris Brown, 2004, age 23, 120 att, 5 games-- Chris Brown had 27 attempts in week 5 and a career high 32 attempts in week 8, before a week 9 bye, came back to have 20 attempts in week 10, then missed next 2 games with turf toe injury. Played in 2 more games, then missed final 3 with same injury. Had missed 5 games previous season with hamstring injury. Brown averaged 4.8 ypc in 2004, and has averaged 3.8 over the last few seasons, and was a non-factor at age 25 last season.
9. James Jackson, 2001, age 25, 119 att, 5 games-- Jackson had 32 attempts in week 2, 26 in week 4, and 24 in week 6. Played two more games after week 7 bye week, before first injury. Suffered ankle injury that forced him to miss 2 games (weeks 10 and 12) and then suffered another season-ending ankle injury early in week 14 vs. JAC. Jackson had a total of 130 more rush attempts over next 4 seasons.
10. Domanick (Davis) Williams, 2005, age 26, 113 att, 5 games--Davis had back to back games of 28 attempts in weeks 8 and 9, then missed next 2 games with a knee injury. Coming back for the then 1-win Texans, Davis had 13 attempts in week 12, then followed with a 3 game stretch of 25, 29 and 22 attempts, before missing the final 3 weeks with further injury to the same knee. Missed all of 2006 season due to same knee injury, questionable if he will ever play again.
11. Jerome Bettis, 2001, age 29, 112 att, 5 games-- Bettis had 94 carries in the 4 games immediately prior to suffering groin injury in week 12, including 29 attempt game in week 9. Bettis missed rest of regular season, came back to try to play in the playoffs that year.
12. Terry Allen, 1998, age 30, 111 att, 6 games-- Allen had already missed a full season to a knee injury earlier in career, and had missed 6 games previous season. Missed 6 games with ankle sprain.
13. Priest Holmes, 2005, age 32, 105 att, 9 games-- Priest had missed 8 games previous year. No games higher than 22 attempts (week 1), as he was sharing carries with Larry Johnson. Injured neck against SD, on a hit from Shawne Merriman. Has not played since.
14. Robert Smith, 1995, age 23, 104 att, 7 games-- Smith never had more than 20 attempts in a game before suffering an ankle injury that caused him to miss 7 of the last 9 games.
15. Terry Allen, 2001, age 33, 99 att, 5 games -- See Terry Allen, 1998, above. Allen missed 5 games with a broken hand.

Three out of 33 players with at least 138 attempts through six games suffered a serious season ending injury (all three knee injuries) before the end of game 8. Only one out of 179 who had between 90 and 137 attempts were out for the season by game 8. That one was a 32 year old back with recent injury history, and the injury was a neck injury. Further, by my count, ten of the fifteen had games with 25 or attempts--and generally multiple games--in close proximity to the first injury. I will count James Jackson as questionable on this issue, as his injury came a little further removed from his highest workload games. Of the four who did not show any significant workload games prior to injury, three were of advanced running back age and had recent prior injury history. Only Robert Smith (1995) could be considered a younger running back who did not show a recent high workload near the time of injury onset.

The majority of running backs, though, survived the period of early heavy usage without suffering an immediate injury. 24 of those 28 running backs in the high carry groups that did survive immediate injury went on to finish the season with 15 or 16 games played. So, if they did not manifest a severe injury soon after the high workload period, it did not seem to have a lingering effect. Although I am not going to focus on performance, I did test to see if these backs were wearing down as a result of early workload. They were not. For backs who played in at least six more games, there was no correlation between early workload, and changes in yards per carry from those six games versus the rest of the season. In fact, the highest group showed a slight increase in ypc the rest of the year.

The success rate for the high carry groups in getting to the end of the season relatively unscathed, if they survived immediate injury, was comparable with their peer group of feature running backs in the 19.0 to 22.9 attempts per game. I also checked games played by backs in the high workload groups against prior workload history (measured by total career attempts divided by career games), raw career attempts totals, and age. Prior workload history did not matter either way--those injuries hit both high and low career workload backs. What mattered was the present workload. Age also showed no correlation, although the 25-27 age group performed better in this regard, and the games missed increased in both the younger and older groups.

Among the lower carry groups (those who had 113 or fewer attempts), the reason for decline in average games played is not serious injuries. These groups tend to have a higher number of aging backs, backs with recent injury concerns, or backs "playing through injury." As a result, a lot more of these backs are missing some games due to nagging injuries--missing a game here, two games there. Workload, in the short term, is indicative of good health, at least as of the time the running back is being subjected to those additional carries.

In the next post, I am going to examine the end of season and playoff workloads, and their effects on injury rates early the next season.

On their face, the two items named in the title are almost identical. In both cases:

1. you pay a relatively small amount of money;

2. you might or might not receive an enormous amount of money at some point in the future;

3. the long term expected value of the investment is negative. In other words, it's not likely that you'll come out ahead on the deal.

So, aside from the fact that the expected value is probably a bit lower on most lotteries, why is an insurance policy considered a sound financial decision while the powerball-ticket-a-day plan is frowned upon? The answer has to do with what a mathematician might call independence of events. In the case of insurance, the receipt of the enormous sum of money is directly tied to some other random event, like your house burning down. If you never get that enormous sum, that just means you never needed it. With the lottery, on the other hand, the payoff is independent of the rest of the events that might impact your financial situation. If you spend your money on lottery tickets instead of insurance premiums, you might end up homeless, or you might end up with a lot of money that won't necessarily make your life any better.

This is why fantasy football owners of LaDainian Tomlinson will also be drafting Michael Turner this year, and overpaying for the privilege. And why that's OK.

If you have Tomlinson, then Turner is an insurance policy. If Turner finishes the year with 65 carries, who cares? That probably means Tomlinson stayed healthy and productive. If Turner rushes for 1200 yards, it's extremely likely that you will be in desperate need of those yards.

If you don't have Tomlinson, then Turner is a lottery ticket. Sure, the upside is that you get a 1200-yard back with your 9th round pick. But if so, you may not even have a place in your lineup for him. And the more likely scenario is that Turner finishes with fewer than 100 rushes and provides no help if and when your top back gets hurt.

This phenomenon has been well known among fantasy football players for a long time, and it even has a name: handcuffing. But I think the same philosophy, to a lesser extent, can be applied more broadly. For instance, I mentioned yesterday that I love the Laurence Maroney / Tom Brady combination this year because of its potential to deliver consistent weekly production. But there is another reason to like this combo: we know (inasmuch as we ever "know" anything) that the Patriots are going to score a ton of points. What we don't know is who is going to score them. I'm not sure I'd be very comfortable with Maroney at his current price --- there is too much of a risk of Belichick opting to really open up the passing game. For similar reasons, I'm not sure I'd be comfortable with Brady at his current price. But I would be comfortable with the pair for their combined price. I wrote something similar about Brandon Jacobs and Eli Manning last month.

In this article from last year, I wrote:

If I ended up drafting running backs, quarterbacks, and/or Antonio Gates in the first few rounds, I would not hesitate to draft both Rod Smith and Javon Walker and pencil them both in as every-week starters.

My footballguys and p-f-r blog colleague Chase subsequently blamed me for telling him to draft Rod Smith. But it's important to realize that I wasn't necessarily recommending Smith, nor was I recommending Walker. True, part of my recommendation was based on the week-to-week consistency I talked about yesterday, but part of it was also based on the handcuffy principles I'm talking about here. Smith turned out to be a wasted pick, but that must at least in part be tied to the fact that Walker was a bargain at his draft slot. It might have turned out the other way, for all I knew at the time, but I was pretty sure that the Denver WR group would produce at least one player --- and possibly two --- with big numbers.

I realize that that particular recommendation doesn't exactly represent a triumph of this mode of thought; it's just an illustration. And I need to give Chase a hard time.

I'll close with some speculation about same-NFL-team running back pairs. I have never run the numbers because I don't think there are enough numbers to run at this point, but my guess is that such pairs --- like Bush/McAllister and Jones/Barber --- show more week-to-week consistency than similar scoring pairs not from the same team. If that's true, then it makes those pairs attractive for two reasons: the consistency, and the handcuff. If you feel like having some fun this year, spend your first picks on Manning and/or Gates and/or stud wide receivers, and then roll with Marion Barber and Julius Jones as your every week starters. There is, of course, the possibility that one of those guys will get hurt or be relegated to a severely diminished role, but in that case, the other will likely outperform his draft position greatly. I'm not saying that's going to get you good production from the RB position, only that it is a very cheap way to guarantee yourself some minimal output at RB while you enjoy an advantage at the other positions.

It seems to me like fantasy football players are becoming more and more concerned about week-to-week consistency. Lots of people are down on Chad Johnson, for instance, because, while his overall numbers were fairly solid last year, he achieved those numbers via a couple of enormous games and a lot of stinkers. Give me someone who puts up the same numbers, or even lesser overall numbers, but gives me consistent production I can count on from week to week, says the pro-consistency crowd.

There are two issues here.

First, does consistency really make your team better? In the very first football article I ever wrote (nine years ago!) I examined that question mathematically, and observed that consistency isn't inherently good or bad. It's good if your team is good and bad if your team is bad. If your team is stronger than your opponent's on a given week, then you just need your players to do their usual thing. You want consistency. If, on the other hand, you have the weaker team, then consistency hurts you. If your guys turn in typical efforts, you're going to lose. You need some unexpectedly big performances. In addition, unexpectedly poor performances don't turn a win into a loss; they just turn a loss into a bigger loss.

In some sense, these are the same reasons why teams that are ahead more often opt for the kneeldown --- the most consistent play in the playbook --- while teams that are behind are more likely to run hook-and-laterals and Hail Marys.

So consistency isn't necessarily a good thing but, since no one plans to have a bad team, it is certainly justifiable for a fantasy owner to, in August, seek to fill his roster with consistent players.

Which leads us to the second issue: how do you find consistent players?

I've studied that issue. A lot. A whole lot. And I always end up concluding that predicting future consistency from past consistency is very difficult if not impossible. That's not to say that consistency can't be predicted. Maybe it can. But if so, there's got to be more to it than just observing which guys were consistent last year or the year before. For example, it could be the case that, because of the Patriot/Brady propensity for spreading the passes around, Randy Moss is likely to be less consistent in 2007 than other non-Patriots who end up with similar 2007 totals. But that's not clear. And other examples like that are hard to come by without introducing a lot of subjectivity and guesswork. For now I'll just say that I'm very skeptical that anyone has any real idea whether Torry Holt, Chad Johnson, or Steve Smith will be more consistent from week to week in 2007.

But there is yet hope. Even if a fantasy GM can't identify consistent players, he can still make his team more consistent --- or at least improve the chances of his team being more consistent --- by acquiring the right combinations of players.

Over the years I've written a few articles on the advisability of having same-team WR/WR pairs, or same-team QB/RB or RB/WR pairs, on your fantasy team (here is one, another appeared in ESPN magazine last year; it used a different methodology but confirmed the conclusions of the linked article.). Those conclusions are that, more often than not, same-team WR/WR, WR/RB, and RB/QB pairs are more consistent than different-team pairs with similar end-of-season totals.

This year, there are a few such pairs that appear to be obtainable in typical serpentine drafts:

• If you're drafting near the end of the first round, then using your first two picks on Chad Johnson and Rudi Johnson might be an option. Early returns indicate that, unless you're in a 14- or 16-team league, you'll probably be overpaying for one of those two if you do that. Of course it goes without saying that you have to like the individual players themselves in those slots for this to be worthwhile.
• If you're drafting near the top of the first round, then Reggie Wayne and Marvin Harrison might be available for your second- and third-round picks.
• One combination I love this year is Chad Johnson and T.J. Houshmandzadeh for the price of a second and fourth round pick.
• Keeping with the Bengals theme, pairing Rudi Johnson and Carson Palmer would be feasible in many leagues.
• Marc Bulger would, in my opinion, be a solid quarterback option for the proud owner of Steven Jackson
• I'm not sure I'm sold on Joseph Addai, but if you are, he and Peyton Manning could be a good combination with your late-first, early-second picks.
• Finally, I'm sure I'm going to hurt my chances of acquiring this combo in my keeper league, but that's how much I care about my readers: my favorite pair this year is Laurence Maroney and Tom Brady. The Pats scored more offensive TDs than all but four teams last season and, despite the fact that Randy Moss is a bit of question mark and Donte Stallworth's and Wes Welker's talents have been greatly exaggerated by many since they signed with the Patriots, there is no doubt that the trio constitutes a major upgrade to the Pats' only offensive weakness. There is very little question that they'll have one of the top offenses in the NFL in 2007. The problem is that nobody knows how Belichick The Inscrutable Genius will choose to score his many, many touchdowns. We don't know if he'll let Brady toss 40 TDs, if he wants Maroney to score 25, or if he's going to mix it up from week to week. If you've got both Brady and Maroney, it's hard to imagine that you won't get solid consistent production every week.

I've been peddling this schtick for awhile now. One objection that I commonly hear --- and it is a valid one --- is that by taking a pair of players from the same NFL team, you are more susceptible to the effects of a single injury. If you have Marvin Harrison and Reggie Wayne, for instance, then you're one Peyton Manning injury away from getting consistently low production instead of consistently high production from your receiver duo.

At least in some cases, that's a price you may have to be willing to pay. If you want to decrease your week-to-week risk, you might have to increase your "macro" level risk. I'll discuss that a bit more in my next post.

Quick post today...

There was some chatter in the comments to Wednesday's post about looking at teams' records against opponents that were better than they were and against opponents that were worse than they were.

Since the merger, there have been 13 teams that did not beat a team that was their superior or lose to a team that was their inferior (as measured, of course, by the simple rating system). As would be expected, most of those teams had either very good or very bad records, with an extreme example being provided by the 1972 Dolphins. This was also mentioned, by DolFan316, in the comments to Wednesday's post.

WOBT stands for Wins Over Better Teams, LAWT means Losses Against Worse Teams, and the final column is the total. Note that home field advantage has been incorporating in deciding who was the "better" team in a given game. For instance, the 1998 Dolphins were considered better than the Bills in Miami, but worse than the Bills in Buffalo. They split with Buffalo that year, but still appear on the list below.

 Tm  Yr     record  WOBT LAWT
=================================
buf 1978    5-11-0    0   0   0
oak 1970    8- 4-2    0   0   0
mia 1972   14- 0-0    0   0   0
det 2001    2-14-0    0   0   0
tam 1976    0-14-0    0   0   0
den 1975    6- 8-0    0   0   0
bal 1982    0- 8-1    0   0   0
sfo 1974    6- 8-0    0   0   0
atl 1998   14- 2-0    0   0   0
mia 1998   10- 6-0    0   0   0
was 1986   12- 4-0    0   0   0
cin 1976   10- 4-0    0   0   0
stl 1975   11- 3-0    0   0   0

Here are the teams that were involved in the most upsets:

 Tm  Yr     record  WOBT LAWT
=================================
was 2000    8- 8-0    6   5  11
phi 1983    5-11-0    5   5  10
cle 2002    9- 7-0    6   4  10
chi 1989    6-10-0    5   5  10
nyj 1999    8- 8-0    5   5  10
nyj 1983    7- 9-0    5   5  10
cin 2003    8- 8-0    6   3   9
nyj 2000    9- 7-0    5   4   9
det 1993   10- 6-0    6   3   9
chi 1993    7- 9-0    4   5   9
oak 1999    8- 8-0    3   6   9
min 1983    8- 8-0    6   3   9
rai 1993   10- 6-0    6   3   9
bal 1980    7- 9-0    5   4   9
sfo 1988   10- 6-0    4   5   9

Finally, here is the 2006 list:

 Tm  Yr     record  WOBT LAWT
=================================
was 2006    5-11-0    4   4   8
sea 2006    9- 7-0    4   3   7
ten 2006    8- 8-0    6   1   7
jax 2006    8- 8-0    0   7   7
stl 2006    8- 8-0    4   3   7
atl 2006    7- 9-0    4   3   7
cin 2006    8- 8-0    4   3   7
sfo 2006    7- 9-0    6   1   7
min 2006    6-10-0    3   3   6
det 2006    3-13-0    2   4   6
den 2006    9- 7-0    3   3   6
nyg 2006    8- 8-0    4   2   6
car 2006    8- 8-0    4   2   6
mia 2006    6-10-0    2   4   6
buf 2006    7- 9-0    3   3   6
hou 2006    6-10-0    4   2   6
nyj 2006   10- 6-0    4   2   6
nwe 2006   12- 4-0    1   4   5
phi 2006   10- 6-0    2   3   5
nor 2006   10- 6-0    1   4   5
gnb 2006    8- 8-0    4   1   5
pit 2006    8- 8-0    1   4   5
dal 2006    9- 7-0    0   5   5
kan 2006    9- 7-0    2   2   4
cle 2006    4-12-0    3   1   4
tam 2006    4-12-0    4   0   4
bal 2006   13- 3-0    0   3   3
ari 2006    5-11-0    2   1   3
ind 2006   12- 4-0    1   2   3
chi 2006   13- 3-0    0   2   2
oak 2006    2-14-0    1   0   1
sdg 2006   14- 2-0    0   1   1

The 2006 Jaguars had what I'd call a strange season. They were 7-4 against above average (according to the simple rating system) teams, and only 1-4 against below average teams, including a 20-point loss against the Texans. Here are their opponents, sorted from best to worst, along with the result of the Jaguars game against that opponent:

   Opp  rating  result
======================
   nwe  +10.2    -3
   ind   +5.9    -7
   ind   +5.9   +27
   dal   +3.7    +7
   pit   +3.4    +9
   phi   +3.4    +7
   buf   +2.2    -3
   nyj   +2.0   +41
   kan   +1.0    -5
   mia   +0.7   +14
   nyg   +0.1   +16
   ten   -1.3    -7
   ten   -1.3   +30
   was   -4.0    -6
   hou   -4.5    -3
   hou   -4.5   -20

The correlation coefficient between the two columns of data is +.21, which is pretty strange if you think about it. A positive correlation means that the two quantities tend to vary together, which means that the Jags did better --- not just relatively better, but absolutely better --- against stronger teams than weaker ones. The correlation of +.21 is, in fact, the fourth-highest such value since the merger. For reasons I can't explain, the top three all occurred in 2003.

2003 Cardinals
   Opp  rating  result
======================
   gnb   +8.1    +7
   bal   +6.3    -8
   stl   +5.9   -24
   stl   +5.9    -3   
   sea   +4.1   -18   
   sea   +4.1   -38
   sfo   +3.1    +3
   sfo   +3.1   -36
   min   +2.9    +1
   dal   -0.5   -17
   car   -0.9    -3
   pit   -1.1   -13
   cin   -2.4    +3
   cle   -2.9   -38
   chi   -3.5   -25
   det   -5.8   -18

2003 Giants
   Opp  rating  result
======================
   nwe   +6.9   -11
   stl   +5.9   +10
   phi   +4.4   -18
   phi   +4.4    -4
   mia   +3.4   -13
   min   +2.9   +12
   tam   +1.6    -6
   nor   -0.3   -38
   dal   -0.5    -3
   dal   -0.5   -16
   nyj   -0.6    +3
   car   -0.9   -13
   buf   -1.0   -17
   was   -5.7   -13
   was   -5.7    +3
   atl   -7.4   -20

2003 Vikings
   Opp  rating  result
======================
   kan   +8.3   +25
   gnb   +8.1    -3
   gnb   +8.1    +5
   stl   +5.9   -31
   den   +5.5    +8
   sea   +4.1   +27
   sfo   +3.1   +28
   chi   -3.5    -3
   chi   -3.5   +11
   oak   -5.5   -10
   det   -5.8   +10
   det   -5.8   +10   
   sdg   -6.8   -14   
   atl   -7.4   +13
   nyg   -8.6   -12
   ari  -12.6    -1

It occurs to me that the correlation coefficient might not be the best way to measure the strangeness of seasons, since it can be overly sensitive to outliers. I'm open to suggestions for other ways to capture what I'm trying to capture.

For comparison's sake, here is the least strange season since the merger --- correlation coefficient -.87 --- which also happened in 2003.

2003 Saints
   Opp  rating  result
======================
   ind   +7.0   -34
   ten   +6.5   -15
   phi   +4.4   -13
   sea   +4.1   -17
   tam   +1.6    -7
   tam   +1.6    +3
   dal   -0.5    +6
   car   -0.9    -3
   car   -0.9    -6
   jax   -2.4    -1
   chi   -3.5    +7
   was   -5.7    +4
   hou   -6.0   +21
   atl   -7.4   +28
   atl   -7.4    +3   
   nyg   -8.6   +38   

The least strange season of 2006 was by the 49ers, who were destroyed by the best three teams they played. They were 1-5 against above average teams and 6-4 against below average teams.

2006  49ers
   Opp  rating  result
======================
   sdg  +10.2   -29
   chi   +7.9   -31
   nor   +4.0   -24
   phi   +3.4   -14
   den   +1.3    +3
   kan   +1.0   -41
   sea   -3.6    +6
   sea   -3.6   +10
   stl   -4.0    +7
   stl   -4.0    -3
   min   -4.1    +6
   gnb   -4.4   -11
   det   -6.4    +6
   ari   -6.9    -6
   ari   -6.9    -7
   oak   -9.6   +14

Other 2006 teams who had strange seasons were the Dolphins, who routed the Pats and Bears but lost to the Packers and Texans, and the Lions. Aside from San Francisco, the least strange 2006s belonged to the Giants and Bengals.

So does this mean anything, or is it just trivia? Not expecting to find anything, I ran a regression of Year N+1 wins versus Year N wins and Year N "strangeness coefficient." The result:

Year N+1 Wins =~ 4.97 + .432*(Year N Wins) + .932*(Year N Strangeness Coefficient)

The last coefficient is statistically significant, which indicates that there might be something real there, especially since the result agrees with intuition. If you're capable of beating good teams, then losing to bad teams seems like something you can figure out how to correct. On the other hand, if you are beating bad teams and losing to good ones, it's not clear that you have much untapped upside. The Jags and Giants were both 8-8 last year, but this formula projects the Jags with 8.6 wins this year and the Giants with 7.8. That's quite a difference.

This is an indicator that I haven't heard discussed much. Intuitively, it seems like it could be a general indicator of future success in all sports, not just NFL football. There's more work to be done, but I think there is potential here for identifying groups of teams that are likely to improve next season, much in the same way that deviation from projected winning percentages --- whether projected from the SRS, or from a pythagorean type procedure, or from something like footballoutsiders' DVOA --- does.

Note from Doug: please join me in welcoming JKL to the "staff." He'll be posting articles regularly, or semi-regularly, or whenever he has the time and inclination to bung something down. Glad to have you on board, JKL. [end of Note From Doug]

The Arizona Cardinals have some reasons for optimism entering the 2007 season. Matt Leinart is entering his second year as a starter, with good offensive weapons in the passing game in Anquan Boldin and Larry Fitzgerald. Adrian Wilson is one of the best defensive players in the league. Gone is the coach who was exactly who we thought he was, Dennis Green. Entering is Ken Whisenhunt, who worked under Bill Cowher, and who brings former Steelers offensive line coach Russ Grimm with him.

However, I am not trying to convince you that the Cardinals will be the surprise team of 2007. Predicting a big turn around for the Cardinals is not exactly safe given their history. The Cardinals have been one of the most consistent teams, and it has not been a good consistency. What I will say is that IF the Cardinals turn things around, there will be another factor that could play a role. The Arizona Cardinals will be playing their second season in University of Phoenix Stadium in 2007, after having played the previous 18 years at Sun Devil Stadium in Tempe. If history is an indicator, the Cardinals will enjoy an above average home field advantage in 2007.

Using the very informative site Stadiums of the NFL, I looked at all teams since the AFL-NFL merger who (1) moved to a new stadium, and (2) were playing in the same metropolitan area as the previous season. I found 28 such cases since 1970. Here are the home and road winning percentages for those teams for a 10 year period that includes the two seasons prior to moving into the new stadium, and the eight seasons afterward. Year N represents the first year playing in the new stadium. The first number is the home winning percentage, the second is the road winning percentage, and the third number is the difference between home and road.

• Year N-2 0.488 0.342 +0.146
• Year N-1 0.549 0.413 +0.136
• Year N 0.596 0.404 +0.192
• Year N+1 0.624 0.397 +0.227
• Year N+2 0.552 0.408 +0.144
• Year N+3 0.639 0.454 +0.185
• Year N+4 0.513 0.463 +0.050
• Year N+5 0.548 0.477 +0.071
• Year N+6 0.604 0.399 +0.205
• Year N+7 0.583 0.417 +0.166

Just so you're reading this correct, Year N+1 is the second year in a new stadium, when the 28 teams won 62.4% of home games and 39.7% of road games for a 22.7% difference, the largest of any year in the period examined. The NFL average home-road difference is going to be closer to 15-17%.

I excluded relocated and expansion teams from this because I wanted to see the effect on home performance, and hoped to minimize the effect of changes in road disadvantage due to a team changing its geographic location and travel schedule, or learning a new league. As we can see, the road performance in Years N-1 (the last year in the old stadium) and N, N+1 and N+2 (the first 3 years in the new stadium) are very similar. The changes were to the home field performance, with an increase in home performance in year 1 of the move, and a further increase in year 2 before beginning to revert back in year 3.

There is a potential explanation for this. It has to do with the familiarity and comfort level of the road team.

From 1986-2005, home teams won 57.2% of divisional games, 58.7% of conference matchups, and 59.8% of inter-conference games. Further, in divisional matchups featuring similar climate rivals from the same time zone, the home team was 340-307 (0.526) for the period 1986-2005. When we look at AFC-NFC matchups between similar climate opponents, where the teams play at each venue less frequently, the home team was 143-101-1 (0.586). (NOTE: Similar climate rivals is defined as teams within 10 degrees of each other's average monthly (Sept. to Dec.) temperature, 72 degrees was used for dome teams, and the monthly average was used for outdoor teams.)

It also makes sense that home field advantage would not be at its strongest immediately upon opening the new stadium, if familiarity plays a role in home field advantage. At the start of year one in a new stadium, the home team is basically no more familiar with the nuances of the new stadium than the visitors. By the start of the year two regular season, the home team will have had at least 12 games experience (preseason and regular season) in the new stadium, while every visiting team will have played 1 or fewer games in the stadium.

Arizona's division rival Seattle presents an interesting recent case study on the potential role of familiarity in home field advantage and road disadvantage. In 2002, Seattle moved in to Qwest Field and changed conferences, moving from the AFC West to the NFC West. In 2003, the Seahawks had a perfect 8-0 record at home, and were 2-6 on the road. When teams traveled to Seattle in 2003, they were almost universally playing in a new situation, as prior to 2000, the Seahawks were a dome team. When Seattle went on the road, they were playing in stadiums they rarely, if ever, had visited in the previous decade. The AFC road opponents that year were Cincinnati and Baltimore, and Seattle had not played a road game in either of those cities since 1997. The conference road opponents were Green Bay, Minnesota, and Washington. Seattle had played at Washington in 2001 and at Green Bay in 1999. Seattle lost all 5 of those road games, and went 2-1 on the road within their new division.

This all leads me to believe that the Cardinals in 2007 will enjoy a better home field advantage. The new stadium alone is not enough to turn a bad team into a playoff team. But if the Cardinals can buck history and improve to an average level, the effects of playing in a new stadium in year 2 could mean the difference between a nice improvement, and a playoff appearance.