Quickly… who is the greatest NFL quarterback when the game is on the line in the fourth quarter?

Most of you said “John Elway.” That’s the safe answer, one the mainstream media would approve. You could probably get away with it in any sports bar across the country.

Just don’t say “Tim Couch.” That would be wrong. The bartender would look at you sadly, shake his head, and refuse to honor any more of your requests for Cosmopolitans and Strawberry Daiquiris.

I’ve been studying quarterback performance recently. I now have a database of more than 20,000 quarterback performances, dating back to 1974. I chose 1974 as a starting point because that was when rules were enacted opening up the passing game. It’s the beginning of the modern NFL.

In my study, I tried to identify quarterbacks truly responsible for wins and losses and fourth-quarter comebacks. From 1996 onward, I have game logs, and can figure out what happened, more or less, during a game.  Prior to 1996, I only have raw statistics, so I had to take a few shortcuts.

I awarded a comeback victory to a quarterback who led a scoring drive in the final eight minutes of the game that gave his team the lead for good. In order to receive full credit, the team had to gain at least one first down during that winning drive. Very roughly, one in four NFL games includes a comeback victory drive. That’s part of the NFL’s charm, and a testament to the parity in the league - even going back prior to the salary cap era.

Before 1996, I awarded a comeback victory for a winning score in the fourth quarter or later. When estimating how many comebacks a player should earn, I factored in that these comebacks were 29% more prevalent than comebacks under the later definition. I also could not award wins and losses as accurately, so I gave the win to the starting quarterback, except when he threw less than 30% of the team’s pass attempts in a game. I included all playoff games in this study.

Counting up total comeback victories is only part of the story here. Of course, the top ten list is a who’s who of great signal-callers.

Quarterback	Comebacks
John Elway	52
Dan Marino	47
Joe Montana	35
Brett Favre	35
Warren Moon	34
Peyton Manning	33
Drew Bledsoe	31
Jim Kelly	31
Dan Fouts	29
Vinny Testaverde	28
Ron Jaworski	28

Of course, this is just part of the story. Favre has 306 career decisions, Marino 257 and Elway 247. This top ten list, in fact, contains all the quarterbacks who have 190 decisions in the last 36 seasons. Longevity puts you on the list more than anything.So I needed to figure out how many comebacks to expect from each quarterback. Part of that is simply related to the number of wins a player participated in, but better teams have a lower percentage of their wins in come-from-behind victories, so I adjusted the expected comeback number based on the number of wins a quarterback’s team averaged.

Finally, I divided the number of comeback wins by the expected number to come up with a clutch rating. 100 is average, meaning the number of comebacks matched expectations. And that’s where we find Tim Couch.

Not much more to tell about this story other than showing the full comeback chart. I included every quarterback with 50 or more decisions since 1974.

Quarterback	Record	Comebacks	Expected Comebacks	Clutch Rating
Tim Couch	21-36-0	11	5.2	213
Don Majkowski	26-29-1	14	7.5	186
Jeff Blake	38-60-0	17	9.9	171
Jay Cutler	25-28-0	10	6.1	164
Aaron Brooks	39-52-0	16	9.8	162
Doug Williams	45-43-1	21	14.2	148
Mike Livingston	16-36-0	7	4.8	145
Marc Wilson	34-32-0	15	10.6	141
Steve Bartkowski	59-68-0	25	18.0	139
Brian Sipe	57-52-0	23	16.9	136
Marc Bulger	41-53-0	13	9.6	135
Gary Danielson	29-31-1	11	8.3	133
Erik Kramer	32-37-0	12	9.1	132
Bob Avellini	25-27-0	10	7.8	128
Jeff George	44-80-0	18	14.2	127
Lynn Dickey	45-57-2	17	13.6	125
Mike Pagel	17-37-1	7	5.6	124
Jake Delhomme	58-41-0	18	14.8	122
Jon Kitna	46-74-0	14	11.5	122
Steve DeBerg	57-94-2	23	19.1	120
Brad Johnson	77-55-0	22	18.4	120
Ben Roethlisberger	68-26-0	18	15.2	118
Drew Bledsoe	101-98-0	31	26.2	118
Steve Beuerlein	48-57-0	16	13.6	118
Jake Plummer	71-72-0	21	17.8	118
Carson Palmer	42-39-0	12	10.3	117
Dan Pastorini	54-38-0	20	17.2	116
Tommy Kramer	54-56-0	19	16.6	115
Chris Miller	32-59-0	12	10.5	114
Jay Schroeder	62-40-0	22	19.5	113
Billy Joe Tolliver	15-37-0	6	5.3	113
Neil Lomax	47-54-2	15	13.3	112
John Elway	161-85-1	52	46.4	112
Trent Green	56-57-0	16	14.4	111
David Woodley	36-18-0	11	9.9	111
Vinny Testaverde	93-117-1	28	25.4	110
Jim Zorn	43-62-0	15	13.7	110
David Garrard	33-32-0	9	8.2	109
Ron Jaworski	79-73-1	28	25.7	109
Ken O’Brien	50-63-1	19	17.5	109
Roger Staubach	68-28-0	21	19.5	108
Dan Fouts	89-81-0	29	27.1	107
David Carr	24-56-0	7	6.7	105
Neil O’Donnell	55-50-0	18	17.2	105
Joe Theismann	82-50-0	27	25.8	105
Richard Todd	50-56-1	15	14.4	104
Peyton Manning	139-66-0	33	31.6	104
Pat Haden	33-22-1	11	10.6	104
Kerry Collins	85-95-0	24	23.2	104
Stan Humphries	51-33-0	14	13.6	103
Warren Moon	105-104-0	34	33.1	103
Dan Marino	154-103-0	47	46.0	102
Joe Ferguson	71-87-0	22	21.5	102
Daunte Culpepper	44-60-0	12	11.8	101
Tom Brady	110-33-0	24	23.8	101
Boomer Esiason	84-91-0	26	25.8	101
Jeff Garcia	58-60-0	15	14.9	101
Steve Grogan	77-61-0	23	22.9	100
Jim Hart	56-54-0	17	16.9	100
Rodney Peete	45-38-0	12	12.1	99
Jason Campbell	19-31-0	5	5.1	98
Randall Cunningham	82-57-1	24	24.4	98
Jim McMahon	70-31-0	19	19.3	98
Archie Manning	25-75-0	9	9.2	98
Bobby Hebert	58-48-0	17	17.7	96
Jeff Hostetler	56-33-1	14	14.7	95
Rick Mirer	24-42-0	7	7.4	95
Bernie Kosar	56-56-1	18	19.0	95
Mark Malone	24-33-0	8	8.5	94
Jay Fiedler	40-24-0	9	9.6	93
Jim Kelly	110-65-0	31	33.2	93
Joe Montana	130-55-0	35	37.7	93
Rich Gannon	77-57-0	19	20.5	93
Doug Flutie	39-33-0	9	9.8	92
Eli Manning	54-39-0	13	14.3	91
Philip Rivers	46-23-0	10	11.0	91
Elvis Grbac	41-35-0	10	11.1	90
Steve McNair	93-66-0	21	23.4	90
Drew Brees	72-54-0	16	18.0	89
Dave Krieg	101-84-0	27	30.7	88
Gus Frerotte	46-49-0	10	11.6	86
Brian Griese	46-38-0	9	10.4	86
Bert Jones	45-47-0	11	12.8	86
Matt Hasselbeck	68-58-0	15	17.6	85
Danny White	66-32-0	16	18.9	85
Ken Stabler	92-51-0	23	27.5	84
Tony Eason	30-25-0	9	10.8	83
Joey Harrington	26-49-0	5	6.1	82
Vince Ferragamo	30-29-0	8	9.8	81
Craig Morton	50-49-0	13	16.1	81
Chris Chandler	68-75-0	14	17.3	81
Bill Kenney	35-41-0	8	10.0	80
Mark Brunell	83-77-0	17	21.3	80
Jim Plunkett	64-44-0	16	20.1	80
Mark Rypien	52-35-0	13	16.4	79
Trent Dilfer	61-55-0	12	15.2	79
Tony Romo	39-20-0	7	8.9	79
Chad Pennington	45-41-0	9	11.5	78
Fran Tarkenton	49-22-2	11	14.2	77
Scott Mitchell	32-39-0	7	9.1	77
Wade Wilson	40-40-0	10	13.3	75
Terry Bradshaw	91-34-0	19	25.4	75
Mike Tomczak	46-34-0	10	13.4	75
Tony Banks	31-44-0	6	8.1	74
Jim Harbaugh	66-69-0	14	19.3	73
Bubby Brister	39-40-0	9	12.5	72
Brett Favre	195-111-0	35	48.5	72
Troy Aikman	103-69-0	21	29.2	72
Jim Everett	66-92-0	15	21.1	71
Donovan McNabb	99-56-1	17	24.2	70
Bob Griese	46-30-0	9	14.0	64
Steve Young	101-50-0	17	27.5	62
Dave M. Brown	23-34-0	4	6.5	62
Phil Simms	99-69-0	20	32.9	61
Ken Anderson	74-63-0	12	22.4	54
Kurt Warner	78-52-0	10	19.6	51
Eric Hipple	28-32-0	4	8.4	48
Kordell Stewart	50-36-0	6	12.5	48
Michael Vick	39-28-1	4	9.0	44

Is Tim Couch the greatest fourth-quarter leader of all-time? I don’t think so. Why not say he’s the worst first-half quarterback of all-time, and came to his senses in the fourth quarter? But when we study clutch performance in any sport, the answers are never what we expect.

Did you know that the “perfect” quarterback rating is 158.33? It was accomplished twice by quarterbacks earning a decision last season - Drew Brees in week 12 against New England (18-23-371-5-0) and Eli Manning in week 5 against Oakland (8-10-173-2-0).

The passer rating, as it’s officially known, was developed almost 40 years ago to provide stats-crazed football fans with a quick look at how effective quarterbacks were in games. But since it’s meant for a full-season evaluation rather than a single-game evaluation, among other issues, it’s not a perfect measure.

For one, it weights four statistics (interception percentage, touchdown percentage, yards per attempt and yards per catch) equally. And second, since it has caps and ceilings over values that can reasonably be achieved in a game, a rather large number of pass attempts is required to get a reasonable rating. For example, go 1-for-1 for 20 yards and a touchdown and you’re considered perfect.

Since I’ve been looking a lot at quarterback performances lately, I wanted a measure that can be used for individual games. I wanted one that weighted statistical components by importance in creating victories. And I wanted a calculation that, while complex, led to something easily understood.

So I created the Quarterback Grade.

We all understand grades. In school, if we did well, we received a bright, bold A. And if we were thoroughly mediocre, we received a C. Teachers would talk about bell curves and grade distributions, and later on, if we were really into mathematics, we’d understand standard deviations and how a distribution curve works. But we did see from the beginning that an A was special, and only a small percentage of the class earned one.

I started out by looking at all the quarterback performances from 1974 through 2009. That’s 36 seasons of data, what I consider “modern” football in that 1974 marked the end of the days when defensive players were encouraged to mug the receivers and having a solid passing game was quite difficult.

I looked at what it meant to win a game. Which statistics correlated most closely with winning. And then I tried to create a formula that reflected those statistics. Some ideas worked better than others. I wanted to include completion percentage in my formula, but found that when combining it with other statistics, it does not improve the accuracy of the formula. Instead, I broke apart completions and incompletions.

In the end, I had a formula that reflected 24% of a win (correlation of 49%). This is 20% more complete than passer ratings. Looking at the 2009 season, the quarterback with the better grade won 83% of his games. That’s compared to 81% for the passer rating. The two measures differed only in 29 games, with the quarterback grade reflecting the winner 17 times.

As a mathematical exercise, that’s nice, but not all that exciting. I believe I’ve added the following benefits as well, though:

The measure is easily attached to the current season. While raw grades are 3% higher today than they were in the ’70s, part of the final grade includes modifying the scale based on the current season (or the previous season if the season isn’t yet over). The passer rating is about 17% higher than it was in the late ’70s. Why the difference? Having more components in the grade, and having larger ranges on the floors and ceilings of individual components.

The letter grades provide an easily understandable quick look at the data. Earn an A, and you’re on the winning team 90.2% of the time. Earn a D, however, and your team is in the win column only 8.5% of the time.

I’ve also attempted to provide some reward for doing more of a good thing. The passer rating is based completely on averages. The quarterback grade is based on raw numbers for interceptions, touchdowns, fumbles, pass completions and incompletions. Only yards per catch and per pass attempt are based on averages. I could have increased the accuracy to about 25% more than the passer rating by purely using averages, but I decided to lose a little accuracy and reward raw numbers a little more.

The formula is as follows:

Raw grade = interception bonus + touchdown bonus + fumble bonus + .7 * (11 - incompletions) (capped/ceilinged at 3 and 24) + .25 * (completions - 17) (capped/ceilinged at 2 and 35) + .58 * (yards per catch - 12.2) (capped/ceilinged at 5 and 24) + 1.1 + (yards per attempt - 6.95) (capped/ceilinged at 2 and 14.4) + year_constant.

The year constant is found by subtracting the average for the year from 77.

The bonuses are based on a chart. For interceptions, for example, it starts with +8 for none, rising to -20 for 5 or more.

Letter grades are assigned based on a normal distribution curve. An A+ is assigned for a score of 100 or more, an A from 92.0 to 99.9, and so on down to failing for a score below 50.6.

Here’s an example of how the scores could look in a table for week 12 of last season (fumbles in parentheses):

Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Arizona	17	Matt Leinart	21-31-220-0-0 (1)	88.1	74.9	C+
Tennessee	20	Vince Young	27-43-387-1-0 (0)	99.7	86.4	B+
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Carolina	6	Jake Delhomme	14-34-130-0-4 (0)	12.8	45.5	F
New York Jets	17	Mark Sanchez	13-17-154-0-1 (0)	79.0	79.9	B-
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Chicago	10	Jay Cutler	18-23-147-1-2 (0)	71.6	73.1	C+
Minnesota	36	Brett Favre	32-48-392-3-0 (0)	112.5	91.5	A-
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Cleveland	7	Brady Quinn	15-34-100-0-0 (0)	51.4	65.1	C-
Cincinnati	16	Carson Palmer	13-24-110-1-0 (0)	80.2	78.2	B-
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Green Bay	34	Aaron Rodgers	28-39-348-3-0 (1)	124.7	90.1	A-
Detroit	12	Matthew Stafford	20-43-213-1-4 (0)	30.5	51.9	D
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Indianapolis	35	Peyton Manning	27-35-244-3-2 (0)	100.2	80.3	B
Houston	27	Matt Schaub	31-42-284-2-2 (1)	87.8	71.0	C
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Jacksonville	3	David Garrard	25-36-307-0-0 (2)	95.5	75.8	C+
San Francisco	20	Alex Smith	27-41-232-2-0 (0)	96.8	83.8	B
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Kansas City	14	Matt Cassel	19-31-178-1-1 (2)	74.4	68.5	C
San Diego	43	Philip Rivers	21-28-317-2-0 (0)	135.6	97.4	A
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Miami	14	Chad Henne	17-31-175-1-3 (0)	42.5	62.5	D+
Buffalo	31	Ryan Fitzpatrick	17-26-246-1-1 (0)	92.8	84.2	B
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
New England	17	Tom Brady	21-36-237-0-2 (0)	55.0	63.6	C-
New Orleans	38	Drew Brees	18-23-371-5-0 (0)	158.3	111.6	A+
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
New York Giants	6	Eli Manning	24-40-230-0-1 (1)	65.6	64.0	C-
Denver	26	Kyle Orton	18-28-245-1-1 (0)	89.1	82.5	B
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Oakland	7	Bruce Gradkowski	18-35-200-1-0 (1)	78.3	72.9	C+
Dallas	24	Tony Romo	18-29-309-2-0 (0)	121.2	94.2	A
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Pittsburgh	17	Dennis Dixon	12-26-145-1-1 (0)	60.6	73.7	C+
Baltimore	20	Joe Flacco	23-35-289-1-0 (2)	100.8	79.4	B-
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Seattle	27	Matt Hasselbeck	14-25-102-0-0 (0)	65.8	72.2	C
St. Louis	17	Kyle Boller	28-46-282-1-2 (0)	67.5	67.0	C
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Tampa Bay	17	Josh Freeman	20-29-250-2-0 (0)	118.5	91.2	A-
Atlanta	20	Chris Redman	23-41-243-2-0 (0)	89.8	81.4	B
Team	Score	Quarterback	Statistics	QB Rate	New Rate	Grade
Washington	24	Jason Campbell	22-37-231-2-2 (0)	73.1	71.0	C
Philadelphia	27	Donovan McNabb	21-35-260-1-1 (0)	80.7	78.2	B-

The grade make it easy to get a general idea of how well the quarterback fared. With the passer rating, we kind of know that 100 or higher is very good, but we don’t have much of a sense of how good. And, while 80 was kinda good 20 years ago, today it’s more mediocre.For full-season comparisons, I like averaging the grades, much like you would if you were giving your class a grade for the entire semester. While grades for partial games could be considered as well, possibly weighting them less, I prefer only to consider games where a quarterback earned a decision.

Here is how Kyle Orton fared in 2009:

Kyle Orton	Result	Grade
at Cincinnati	W, 12-7	B+
Cleveland	W, 27-6	B
at Oakland	W, 23-3	B
Dallas	W, 17-10	B+
New England	W, 20-17	B
at San Diego	W, 34-23	A-
at Baltimore	L, 7-30	C
Pittsburgh	L, 10-28	D
New York Giants	W, 26-6	B
at Kansas City	W, 44-13	C+
at Indianapolis	L, 16-28	B
Oakland	L, 19-20	B-
at Philadelphia	L, 27-30	B-
Kansas City	L, 24-44	C-
14 Decisions	8-6	2.7 GPA (B-)

Note that Denver’s Orton started out with Bs and B+s, but faded in the second half of the season. That’s readily apparent just from looking at the grades.I’m thinking of integrating the quarterback grade into future versions of Front Office Football (if there ever is such a product). This could add a lot to the game rather easily.

The words struck terror in those who follow college football. In December, the Big Ten made it known it wants a twelfth team. The NCAA won’t allow conferences with less than twelve teams to stage a conference championship game. And, as Joe Paterno has said more than once, the Big Ten becomes irrelevant after the last weekend of the regular season. For two weeks, all eyes are everywhere except the Big Ten.

It’s also aesthetically pleasing to have twelve teams in a league, rather than eleven. As long as there’s a way to preserve the annual Michigan/Ohio State rivalry (OK, not lately, but once Rich Rodriguez is a distant memory, we’re confident Michigan will compete again).

When the Big Ten expands, another league will have to replace the lost school. And once this happens, four major conferences will have twelve teams. The fifth, the Pac Ten, announced just last month that it might soon add two teams, so that it will also have twelve. The Big East, more a basketball conference than a football conference, only has eight schools playing football. And because it’s definitely the weakest of the majors, and adding four schools playing football to its already-huge collection of elite basketball teams would be nearly impossible, I don’t see the Big East playing twelve-team football in the next decade.

So, with expansion talk comes candidate talk. And that’s the purpose of this article. How will  schools make these selections? Which are the best candidates? What might happen over the next few years?

First, a discussion of the conferences is necessary. A lot of the discussion boils down to money. And money, these days, comes from television contracts and filling football stadiums.

SEC schools receive about $17 million per year, and Big Ten schools receive about $16 million per year from television revenue. This places them firmly atop the leaderboard. The ACC, Pac Ten and Big XII each earn about $6 million per year. That’s a huge difference, and it’s why those three conferences are all talking about forming their own networks, or combining to form a network.

The ACC, in particular, is disappointed. As adding Virginia Tech, Miami and Boston College was all about adding big cities and television ratings. This hasn’t manifested in more revenue. That’s why, in this discussion, I’m not going to place as much emphasis on large television markets as some. Yes, it matters. But it’s not automatic. The ACC example with Boston shows us expansion outside the conference’s geographic footprint may not be the wisest move.

The television revenue argument explains why a Big XII school might consider moving to the Big Ten, where there’s already much more revenue. Or to the Pac Ten, where adding two schools will result in a championship game and more revenue - a bigger Pac Ten might bridge some of the gap between the lesser majors and the top two.

There is some talk about 14- and 16-team football conferences. I don’t see how that will work, though. It means schedules where you don’t even see a non-division rival visiting more than once or twice a decade. And diluting television revenue. The argument is completely based on expanding into rival conferences’ media bases. The ACC couldn’t even accomplish this with no rivals in Boston. In the late ’90s, the WAC experimented with 16 teams. No one seemed to like the result, and the experiment ended after four years with the creation of the Mountain West.

So I will assume the ACC and the SEC are happy where they are at twelve schools. And that they won’t lose any schools to other conferences. The SEC, as the revenue leader, is very stable. The ACC is less stable, and Maryland might be a great fit for the Big Ten in many ways, but I don’t see Maryland fans being too happy with losing some great ACC rivalries. The ACC might be struggling a bit financially, but it’s still an elite conference.

The Big XII is another story in that its geography also fits the Big Ten and the Pac Ten in places. I can see its north division splintering. The Big XII also would have an easier time replacing schools, as the breakup of the old SWAC left the region with some strong traditional schools having to accept a demotion.

The Big East trails in revenue, and there’s no potential for additional football money. Even with a $2 million penalty for leaving the conference, it would be incredibly lucrative for a Big East school to jump to a larger conference. Since the ACC and SEC are full, that gives the Big Ten choices for expansion.

The Mountain West and Conference USA provide about $1 million per school per year in television revenue. That’s about one fourth of the Big East. So there’s no question candidates from those conferences would be willing to make a move. The WAC has a very weak television contract, and the two conferences at the bottom of the FBS totem pole - the MAC and the Sun Belt - provide little to no television revenue.

When choosing a new school, conferences have varied criteria. Some conferences value academic reputation more than others. Some care about media markets. Some care about traditional rivalries. Academics cannot be overlooked, however, in any case.

If you examine the largest conferences, they are full of schools with top academic reputations. A major school likely has an international reputation as a research institution. There are only a handful of major schools that aren’t listed in the QS World Rankings of 620 universities around the world. Only eight universities among the 66 considered majors do not appear on that list. And only 13 of the 54 mid-majors do appear on that list.

The world rankings have nothing to do with athletics. Conferences were originally formed with academics in mind. Nowhere outside the Ivy League is that stronger than with the Big Ten. The Big Ten is the only major conference placing all its schools in the QS top 250. All eleven conference schools are also members of the Association of American Universities (AAU), an association of 62 research institutions that produce nearly half of all Nobel Prize winners, as one example.

The Big Ten will undoubtedly chose a new member that fits that profile. Only two schools in the general geographic area - Pittsburgh and Rutgers  from the Big East - look academically like the Big Ten schools. Of the two, Pittsburgh is stronger academically, and Rutgers is stronger in terms of bringing in more television viewers - which may or may not manifest given the Boston College example. Another Rutgers advantage is gaining a conference foothold in New Jersey - which has an enormous pool of talented high school football players, many of whom leave the area for major colleges.

I can’t see the Big Ten choosing another school, even though many people are talking about Missouri and Nebraska from the Big XII. They aren’t quite as strong in the academic rankings (both are ranked 400-500 by QS - far behind any current conference school) and it might be hard to convince an original Big 8 team to leave its rivalries behind.

So I’m assuming the Big Ten will take a school from the Big East. And the Big East will be down to seven football schools. The Big East will have to add at least one football school due to NCAA rules requiring eight in a conference. The Big East can definitely take any mid-major in its region, with the draw being its superior basketball teams. In fact, for that reason alone, Memphis from Conference USA would be the obvious candidate.

The Northeast is severely under-represented among major schools. The reason being that the elite academic universities formed the Ivy League. And the Ivies decided decades ago not to offer sports scholarships and not to play, once Division I split, FBS football. Aside from Memphis, which is not on the QS list, one outside candidate - and it’s not a traditional sports power in the slightest - is Buffalo. The Bulls only moved into the FBS a decade ago, but Buffalo (now in the MAC) is the only QS list and AAU university in that region not already in a major conference. Some day, I expect to see Buffalo in the Big East.

The next move comes from the Pac Ten, which is second to the Big Ten in valuing academics. Six of the ten schools are ranked in the QS top 250, none aren’t on the list. Seven are AAU schools. The obvious first choice for the Pac Ten is Utah from the Mountain West. Utah fits academically and has been fielding good sports teams for a while now.

The second choice is a little more difficult. The Pac Ten could try to take Colorado from the Big XII. The Denver market is considered valuable. Colorado is an excellent academic school, and a good fit for the league. But the Buffaloes are also an original Big 8 school and, financially, there might be no advantage to leaving traditional rivals.

If the Pac Ten considers a jump all the way to eastern Texas,  which might be too much of a stretch, one huge prize is Rice, located in Houston. Rice is by far the strongest school academically among the mid-majors. Rice was part of the SWAC, and it’s hard to believe the Big XII chose Baylor instead. But the road trip from Seattle to Houston is more than 2,000 miles. That might be too far, even for a major school.

For the same reason, only more so, another excellent university, Hawaii, is a long shot for the Pac Ten. Hawaii would be a good academic fit for the Pac Ten. Travel issues make the choice a long shot. Other possibilities include Texas Christian and Southern Methodist, located in the Dallas market, New Mexico, Brigham Young, and Tulsa.

If I had to guess, I’d say the Pac Ten will make a big play for Colorado. And if that’s not possible, then Rice will get the call.

If the Pac Ten steals Colorado away, the Big XII will need a new school. The Big XII has only three schools in the top 250, but seven conference members are part of the AAU. Taking Rice would be a logical move. TCU and SMU are possible additions. Another good choice would be Tulane, an AAU school in New Orleans. Louisiana is the only state that has a larger percentage of elite high school players who have to leave to play major college football than New Jersey.

The Pac Ten and Big Ten expansions will leave the Mountain West and Conference USA three teams short. And, down the road, the Mountain West might want to go to twelve schools so it can hold a conference championship as well. So that’s six schools total. Since these conferences are looking at schools with more regional academic reputations, the field is more open.

What will happen? The WAC will get raided. Boise State, Idaho, Nevada and Fresno State could well join the Mountain West. Louisiana Tech and New Mexico State might go to Conference USA. The WAC will probably have to disband, leaving San Jose State, Hawaii and Utah State without a conference, or trying to get into the Sun Belt.

This is a lot of speculation, of course. But given the current emphasis on twelve-team conferences, this kind of movement seems inevitable. And there’s always the possibility of a real post-season tournament - which most fans want desperately - making the conference championship games less relevant and expansion less lucrative. Only time will tell.

In my off-season football studies, I’ve been looking at the numbers quarterbacks produce. I’ve created a database containing every quarterback performance going back 36 seasons.

Why 36? I consider 1974 the beginning of the modern NFL. Prior to that season, it was a rushing league. In 1974, just four years after the AFL/NFL merger, Pete Rozelle and others realized the game needed to open up a little. I believe he saw the potential for the NFL to ecplise the popularity of college football. If only the best players in the game could put on a great show.

In 1974, the rules changed drastically. Overtime was introduced for regular-season games. The goalposts were moved to the back of the end zone, reducing the importance of field goals. Kickoffs were moved back. And, most importantly, restrictions were placed on blocking and contact with receivers.

All of a sudden, the passing game opened up. Quarterbacks became game managers rather than gunslingers. In 1977, blocking and receiver contact rules were revised, and in 1978, the NFL created the modern 16-game schedule and ended defensive contact with receivers past the five-yard mark.

Because of this emphasis, I feel any study before 1974 is comparing apples to oranges. I could easily start with 1978, because it took four years for most coaches to begin adjusting to the new world. But I went with 1974 to show the modern trends developing.

The first chart shows the average number of passing attempts for a quarterback with a decision. I assigned wins and losses to quarterbacks. Since easily-found game logs only go back to 1996, I assigned the decision to the starter from 1974-1995, unless he had less than 60% of the pass attempts of a quarterback who came into the game later. After 1995, I assigned the win or loss to the quarterback who was in the game when the winning score occurred.

Click on any image to view the image at full size.

You can see the effect the rules changes had on passing numbers. Coaches quickly made the forward pass more a part of the game plan. And this added five passes per team per game. Since then, there’s been a slight upward trend in passing, but the average over the last 30 years, 30.4 passes per team, is only about one pass per game higher than we saw in the early 1980s.

The second chart shows the completion percentage by quarterbacks who earned a decision.

The rules changes made a less dramatic difference. An increase of one completion every 25 passes. But, combined with the increased number of throws, you can see how quarterbacks became a larger part of the game. The trend upward in completion percentage is a bit steeper more recently.

You can see the influx of “west coast” passing and safer throws around the beginning of the 1990s. Even in the last ten years, completion percentage is increasing. Some might claim the renewed “emphasis” on the contact rule in 2004 (otherwise known as the Patriot Rule) is responsible for this trend.

In 2009, quarterbacks set an NFL record for most passing yardage. Quarterbacks earning a decision threw for almost 224 yards per game - a far cry from the 153 yards per game in 1977.

The third chart shows yardage per catch.

What we see here is an illustration of the “west coast” effect on game planning much more dramatically. Teams stopped firing the deep ball so frequently when they did pass. Possession receivers became far more valuable. Obviously, increases in completion percentage have gone hand-in-hand with decreases in yardage per catch since the 1970s.

This means yardage per passing attempt has remained relatively static over the years.

There was a small peak in the early 1980s, but, since 1978, the average yards per pass attempt has fluctuated only between 6.7 and 7.2.

And we see a very similar trend in the percentage of passes thrown for touchdowns. A peak in the early 1980s, but very little fluctuation over the years.

The last chart is perhaps the most dramatic in the series.

This explains a lot. Turnovers make a huge difference in football. Quarterbacks have reduced their interception percentage more than 30% over the last 35 years. “West coast” has made a difference here, but the numbers have steadily trended down regardless. This more than makes up for tiny declines in yards per attempt and touchdown pass percentage since the early 1980s.

A reduction of what amounts to about seven inches per pass attempt seems well worth about .3 fewer interceptions per contest. That, combined with the higher completion percentage, keeps the chains moving more steadily.

As labor negotiations between the NFL owners and the players’ association continue to move as slowly as a 340-pound lineman in a 40-yard dash, it’s becoming more and more apparent that the salary cap era has come to a close.

In 1994, the players and owners agreed on a deal that limits each team in spending. But, just as importantly, it forces teams to spend a minimum percentage of the cap. This deal ushered in 16 seasons of relative labor peace. As well as a certain measure of parity. While it didn’t change the difference between the top teams and the bottom teams in a given year, it did greatly lower the number of years required to turn a bad team into a good team.

And it also forced teams to employ good general managers if they wanted to compete. Who can forget the mess the Detroit Lions have made, simply by not having good football people in the front office?

So, in tribute to the Salary Cap Era, I wanted to post a list of the franchises, in order of the wins each accumulated from 1994 through 2009. I included playoff victories. I also separated franchises by home city. So the Tennessee Oilers/Titans are under Tennessee, while the Houston Oilers and Houston Texans are considered the same franchise. This might make the list more relevant to each team’s fan base.

The (*) next to a franchise indicates the number of years there was no team in that city during the last 16 seasons.

Franchise	Wins
New England	183
Pittsburgh	175
Green Bay	171
Indianapolis	171
Denver	161
Philadelphia	155
Minnesota	151
Dallas	146
San Francisco	140
New York (NFC)	139
Miami	138
San Diego	135
Seattle	132
Tampa Bay	132
Jacksonville	130 (*1)
Kansas City	130
Tennessee	125 (*3)
Baltimore	124 (*2)
Atlanta	124
New York (AFC)	124
Carolina	123 (*1)
Chicago	123
Buffalo	121
New Orleans	120
Washington	114
St. Louis	112 (*1)
Arizona	105
Oakland	101 (*1)
Cincinnati	100
Detroit	88
Cleveland	76 (*3)
Houston	66 (*5)
Los Angeles (AFC)	9 (*15)
Los Angeles (NFC)	4 (*15)

Go west, young football player. But don’t come back. That’s the crystal-clear message I found as I looked through the last 14 seasons of NFL game data.

I wanted to examine the NFL road trip. One of the great mysteries of professional sports is the poor performance of road teams. Especially in the NFL, where every team has a week to recover from each game. Unlike the NHL, with its rules about line changes and putting your stick on the ice for a faceoff, the NFL gives no direct advantage to the home team. Unlike baseball, where the home team bats last every inning, the NFL tosses coins and gives each team equal opportunity to face each goal. They even let visiting teams wear their cooler white jerseys on the road, which could mean a tiny, tiny bit on a warm September Sunday in Phoenix.

Yet visiting teams have won only 41.4% of the time in the NFL since 1996. That’s a staggering 313 games under .500. Not even the Detroit Lions could approach that mark of futility. At least not since Matt Millen was shown the door.

What’s going on? Is it the crowd? Evidence suggests crowd noise leads to far less than one five-yard penalty per game. Psychologists might say athletes perform better if they are rewarded with cheers. Unfortunately, there’s no game data available measuring crowd reaction.

The raw numbers suggest the difference is subtle. The average line from a quarterback who earned the decision as a visitor is 18.2-31.1-209.0-1.2-1.0. That’s 58.6% completions and an 11.5-yard per-catch average. The average line from the home-team quarterback is 18.3-30.6-212.7-1.3-0.9. That’s 59.8% completions and an 11.6-yard per-catch average. These are essentially equal, especially when you consider that the team that’s behind late in the game will throw more passes, and riskier passes.

I took an especially close look at games where one team overcame a deficit in the last eight minutes of the contest. Where a quarterback led a game-winning drive. There were 873 such games in the last 14 years. This accounts for 27.4% of all visiting-team wins and 21.6% of all home-team wins.

Yes, when the game is close and every drive counts, home field means very little. The visiting team wins 47.3% of those games. A statistically significant difference over 14 years, but a much smaller one. In fact, when visiting teams win, they win by less of a margin.

Visiting teams perform well enough in close games. Looking at margin of victory as a percentage of all wins, visiting teams, for example, win by three or less points in 24.9% of all their victories. Home teams win by three or less in 21.9% of all their victories. The difference is consistent at every margin up to ten points. At ten points, the percentages switch to the home team, where it remains. Home teams win by 16 or more points in 32.6% of their wins. For visiting teams, that number is 25.0%.

Yet, intuitively, we’d say the crowd would be louder and more a factor in close games. So, why else would a team perform badly on the road? There could be an advantage to knowing a field better than your opponent. Where will the wind affect a pass? What will the turf allow you to do? Some fields have a bigger crown than others.

And then there’s the travel. The wear and tear of getting onto an airplane and going to another city before a game.

So I looked at the distance a visiting team had to travel before a game. If simply getting out of town is a factor, then maybe a longer trip would be a bigger factor.

It is. Teams traveling less than 200 miles for a road game win 46.6% of the time. This gradually decreases until teams are traveling 1400 or more miles. Those teams win only 37.1% of the time. In other words, the distance traveled can multiply home-field advantage by a factor of almost four.

But it’s not that simple. Not at all. Because the breakdown of this factor suggests distance is only part of the solution. Teams traveling west 1400 or more miles win 41.3% of the time - almost exactly the numbers you’d see from an average road team. But teams traveling east 1400 or more miles win only 32.6% of the time. Now we’re up to a factor of more than five.

I looked at the teams involved, as only eight different teams can make an eastbound road trip of that distance (two of which - Houston and Dallas - have made only seven such trips in the last 14 seasons combined). Looking at the other six, it’s a good cross-section of the NFL. Denver (32.0%) averaged 10.4 wins per season (the NFL average, including playoff games, is 8.5). San Diego (37.9%), San Francisco (32.3%) and Seattle (36.0%) were fairly average teams. And Arizona (26.6%) and Oakland (32.0%) were significantly worse than average. This is a real trend.

But why westbound and not eastbound?

The NFL doesn’t seem very interested in changing starting game times based on the visiting team. A team traveling west across three time zones will start its game at 4:00 local time. And a team traveling east across three time zones will start at 10:00 a.m. local time.

Perhaps an athlete needs a certain amount of time after he wakes up to prepare his body for play. Perhaps he loses sleep when he has to wake up three hours earlier.

So I further broke down visiting team travels, this time by time zone. And yes, I even factored in the facts that Phoenix switches to Mountain time in November (October until 2007) and Indianapolis used Central time through October until 2006. And what I found was surprising.

Visiting teams staying in their own time zone won 43.9% of the time over the last 14 years. Visiting teams traveling west one or more time zone won 41.4% of the time (one time zone, 42.0%; two zones, 39.7%; three zones 41.5%). Distance seems to have a small effect, perhaps related to the interruption in routine of the travel itself.

However, traveling east produced more pronounced effects.  Traveling east one time zone produced wins 39.9% of the time. Teams traveling east two time zones won 36.3% of the time and traveling east three time zones gave teams a 33.5% success rate. While I can’t break out night games (when presumably players are not sleep-deprived) all that easily, I would suspect (and sample size would be awfully small here) the dramatic slope of these results would be reduced.

Because of these numbers, I believe the NFL makes life too difficult on teams outside of the Eastern time zone, and I would recommend that all road games start at 1:00 in the visiting team’s time zone, not the home team’s. Barring that, I would recommend that coaches of teams playing in the west adopt an extreme early-bird schedule during the season, except in weeks when they’re hosting a night game.

Sometimes, a random investigation of statistics can lead to important changes in Front Office Football. Going back to the first edition of the game, fumble recoveries have been a bit off. While the total fumble numbers mirror the NFL very closely, as well as who recovers the ball, what happens after the recovery isn’t as well studied.

I had that in the back of my mind a couple of times over the years, but never did the legwork. A couple of weeks ago, as I was watching the bizarre overtime finish between Arizona and Green Bay, I decided it was time to take a good luck at the numbers.

I happen to have the 2007 season boxscores in an easily searchable format, so I started my investigation with that year. What I saw was enough to warrant some changes. There were 27 fumble return touchdowns in the NFL in 2007 - all by the defense. One was on a fumbled punt snap, one on a kickoff return. Only three were on designed run plays, and six were on completed passes.

The surprise? Other than the lack of returns on run plays was the fact that 16 of the 27 touchdowns came on sacks or plays where the quarterback bobbled the snap. So it’s time to reflect this in the game.

The 2010 NFL Amateur Draft Order:

1. St. Louis (1-15) 133 opponent wins
2. Detroit (2-14) 134
3. Tampa Bay (3-13) 142
4. Washington (4-12) 126
5. Kansas City (4-12) 132
6. Seattle (5-11) 122
7. Cleveland (5-11) 131
8. Oakland (5-11) 135
9. Buffalo (6-10) 132
10 (t). Jacksonville (7-9) 127 (subject to coin flip)
10 (t). Chicago (7-9) 127 (subject to coin flip)
12. Miami (7-9) 143
13. San Francisco (8-8) 122
14. Denver (8-8) 135
15. New York Giants (8-8) 137
16 (t). Carolina (8-8) 138 (subject to coin flip)
16 (t). Tennessee (8-8) 138 (subject to coin flip)
18. Pittsburgh (9-7) 125
19 (t). Atlanta (9-7) 129 (subject to coin flip)
19 (t). Houston (9-7) 129 (subject to coin flip)
21. New York Jets (9-7)* 132
22. Baltimore (9-7)* 134
23. Arizona (10-6)* 114
24. Cincinnati (10-6)* 126
25. New England (10-6)* 132
26. Green Bay (11-5)* 113
27. Philadelphia (11-5)* 124
28. Dallas (11-5)* 125
29. Minnesota (12-4)* 113
30. New Orleans (13-3)* 109
31. San Diego (13-3)* 116
32. Indianapolis (14-2)* 121

* - playoff teams reseeded based on how far they go in the playoffs.

Taking a quick look at the top eleven picks in the 2010 draft based on the early results:

1. St. Louis 1-15
2. Detroit 2-14
3. Tampa Bay 3-13 (could be 3 or 4, depending on Kansas City)
4. Kansas City 3-12 (3 with a loss today, 4 or 5 with a win)
5. Washington 4-11 (4 or 5 with a loss today, 5 or 6 with a win)
6. Cleveland 5-11 (could be 6 or 7, depending on Seattle)
7. Seattle 5-10 (5 or 6 with a loss today, 7 or 8 with a win)
8. Oakland 5-10 (7 or 8 with a loss today, 9 with a win)
9. Buffalo 6-10 (could be 8 or 9, depending on Oakland)
10. Jacksonville 7-9 (could be 10 or 11, depending on schedule strength)
11. Chicago 7-9 (could be 10 or 11, depending on schedule strength)

With one Sunday remaining in the season, here’s a look at the top ten in next spring’s amateur draft:

Order	Team	Record	Opponents	Range
1	St. Louis	1-14	126-114	1-2
2	Detroit	2-13	124-116	1-3
3	Kansas City	3-12	125-115	2-6
4	Tampa Bay	3-12	134-106	3-6
5	Washington	4-11	120-120	3-8
6	Cleveland	4-11	121-119	3-8
7	Seattle	5-10	114-126	5-10
8	Buffalo	5-10	124-116	5-10
9	Oakland	5-10	125-115	7-10
10	Chicago	6-9	118-122	7-12

The range above refers to the highest or lowest a team can finish in the draft order.

I want to examine the race for the top pick a little more thoroughly. It will belong to the Rams should they lose to San Francisco, a one-touchdown favorite. But what if Detroit loses to Chicago and the Rams win?

There can’t be a coin-flip for this pick, because St. Louis beat Detroit, head-to-head, for its only win to date. This means the Rams would need a lower opponent strength-of-schedule than the Lions to receive the first pick.

The race would be razor-thin, with eight different game affecting strength of schedule.

The Lions would have a one-game deficit (which is good in this case), gaining from Pittsburgh (Miami), Minnesota (New York Giants), Cleveland (Jacksonville), Green Bay (Arizona), Cincinnati (New York Jets) and Baltimore (Oakland).

The Rams would gain from Jacksonville (Cleveland), Indianapolis (Buffalo), Houston (New England) and Arizona (Green Bay).

If the odds-on favorites win each of these games, and St. Louis wins while Detroit loses, then St. Louis would have a schedule strength of 133-123 and Detroit a strength of 134-122, giving the first pick to the Rams. However, Cleveland is favored by one point over Jacksonville. Reverse that result and strength of schedule result above reverses as well. So this race is much, much too close to call.