I apologize about the length, but I'm excited about this and hope you find it interesting: Feel free to skip to the good stuff (projections) at the end.
I'm posting a preseason ranking based on a model I have been developing with @SunDance Ag over the last several years. We have a preseason model that projects outcomes better than Vegas and better than any other model I have seen out there. We outperform Sagarin, Football Outsiders, and just about everyone else you can find by a significant margin whether you're talking about predicting point spreads or game outcomes. In fact, our model based only on preseason information even beats most reputable models who update their information throughout the year.
@SunDance Ag (Steve Wiggins) and I (Brian Segers) are both economists. I work in consulting and regularly use sophisticated econometric modeling and tools for merger analysis and complex litigation issues. @SunDance Ag is a professor of economics at A&M. We hope to publish this model in a quality journal at some point.
The analysis below is built on a system for forecasting college football team quality. Our system projected us to be very successful in 2012 in the SEC, against most other predictions nationwide. In 2014 it was also very close to pegging our final ranking. The great news for this season is our model is far more optimistic about A&M than the pundits would have you believe and our ranking would indicate.
This system is based on predicting the line for each game—similar to Sagarin and other similar ratings for those of you familiar with those. Since this is a line prediction system using team quality, your first question should be how does this system stack up against the king of lines—Las Vegas. Since the system is a preseason system, the comparison standard is the Vegas preseason lines.
Over the last two seasons, this system consistently beats the Vegas preseason lines (those that were around beginning in July). In the 151 games where there was a difference of 5 or more points between our prediction and the Vegas line, we were correct on 97 (64 percent). For all 426 games with lines in July, this system beat Vegas 241 times (56.6 percent). In other words, this is a better ranking than Vegas right now.
The system is based on a database that includes data on CFB Team Performance, returning proven player performance (Roster Strength), and Recruiting (recruit ranking and stars as measured by the 247 composite). We also have a good system for capturing roster changes. Using our model, below are our Top 25 rankings. The rankings are based on team quality. For each team we also present expected wins and conference wins. I may post some of our program strength, roster strength, and and recruiting scores later.
Note that A&M projects at #7, and is forecast at #2 in the SEC West. We move from slightly behind LSU in the pure team quality to higher expected wins becomes of our favorable schedule. Auburn comes in at 21 because of their heavy losses on offense and is the lowest rated team in the SEC West. I believe this is because most people highly underestimate the negative impact by losing so many receivers and their qb at the same time.
The high ranking of all the teams in the West translates into something that might surprise some; all teams in the SEC West have expected wins of less than 10. The reason is the brutal gauntlet of the West. Second, remember that there is uncertainty in these rankings. The actual success of A&M this year will depend on many factors, including close wins, turnovers, injuries, and the like. I am very confident that we will have a good year this season.
A couple of other teams of interest:
Arizona State #26 (we should win by a little more than 10)
Texas #51—should be a long year in Austin (expected wins: 4.6)
TCU #3 should handily beat Minnesota today (favored by 15.2 - 93.6% chance of winning)
-- If Doctson is out that is a big loss for TCU and their odds of winning would drop by around 10 percent.
I apologize for the messiness. I don't know how to create a clean table or I might include other data.
Wins listed below excluded bowl/playoff games. We use a home field advantage of 3 points:
Rank Team Conf. Exp. Wins Conf. Wins
1 Ohio State Big Ten 11.7 7.8
2 Alabama SEC 9.7 5.8
3 TCU Big 12 10.7 7.7
4 Georgia SEC 10.2 6.4
5 Oregon Pac-12 10.9 8.2
6 LSU SEC 9.2 5.2
7 Texas A&M SEC 9.3 5.5
8 Baylor Big 12 10.2 7.2
9 Ole Miss SEC 8.4 4.5
10 UCLA Pac-12 10.1 7.2
11 Oklahoma Big 12 9.0 6.6
12 Arkansas SEC 7.6 3.8
13 Tennessee SEC 8.5 4.9
14 Notre Dame Ind. 9.8 0.0
15 Clemson ACC 9.9 6.7
16 Florida SEC 8.3 4.6
17 Mississippi State SEC 7.5 3.7
18 Michigan State Big Ten 8.8 5.7
19 Florida State ACC 9.1 5.8
20 USC Pac-12 7.6 5.5
21 Auburn SEC 6.2 2.5
22 Nebraska Big Ten 9.1 5.7
23 Georgia Tech ACC 7.4 5.0
24 Arizona Pac-12 8.5 5.7
25 Wisconsin Big Ten 8.8 5.7
SEC West Exp. Wins Conf. Wins Div. Rank .
2 Alabama 9.7 5.8 1
7 Texas A&M 9.3 5.5 2
6 LSU 9.2 5.2 3
9 Ole Miss 8.4 4.5 4
12 Arkansas 7.6 3.8 5
17 Mississippi State 7.5 3.7 6
21 Auburn 6.2 2.5 7
Our Schedule and the probability of winning each game:
Opponent location Margin Probability of Winning
Arizona State Neutral 10.6 0.803
Ball State Home 32.1 1.000
Nevada Home 29.1 1.000
Arkansas Neutral 4.0 0.668
Mississippi State Home 8.8 0.784
Alabama Home -1.5 0.463
Ole Miss Away -0.3 0.500
South Carolina Home 14.9 0.933
Auburn Home 11.8 0.823
Western Caro. Home 1.000
Vanderbilt Away 15.9 0.938
LSU Away -3.7 0.342
I'm posting a preseason ranking based on a model I have been developing with @SunDance Ag over the last several years. We have a preseason model that projects outcomes better than Vegas and better than any other model I have seen out there. We outperform Sagarin, Football Outsiders, and just about everyone else you can find by a significant margin whether you're talking about predicting point spreads or game outcomes. In fact, our model based only on preseason information even beats most reputable models who update their information throughout the year.
@SunDance Ag (Steve Wiggins) and I (Brian Segers) are both economists. I work in consulting and regularly use sophisticated econometric modeling and tools for merger analysis and complex litigation issues. @SunDance Ag is a professor of economics at A&M. We hope to publish this model in a quality journal at some point.
The analysis below is built on a system for forecasting college football team quality. Our system projected us to be very successful in 2012 in the SEC, against most other predictions nationwide. In 2014 it was also very close to pegging our final ranking. The great news for this season is our model is far more optimistic about A&M than the pundits would have you believe and our ranking would indicate.
This system is based on predicting the line for each game—similar to Sagarin and other similar ratings for those of you familiar with those. Since this is a line prediction system using team quality, your first question should be how does this system stack up against the king of lines—Las Vegas. Since the system is a preseason system, the comparison standard is the Vegas preseason lines.
Over the last two seasons, this system consistently beats the Vegas preseason lines (those that were around beginning in July). In the 151 games where there was a difference of 5 or more points between our prediction and the Vegas line, we were correct on 97 (64 percent). For all 426 games with lines in July, this system beat Vegas 241 times (56.6 percent). In other words, this is a better ranking than Vegas right now.
The system is based on a database that includes data on CFB Team Performance, returning proven player performance (Roster Strength), and Recruiting (recruit ranking and stars as measured by the 247 composite). We also have a good system for capturing roster changes. Using our model, below are our Top 25 rankings. The rankings are based on team quality. For each team we also present expected wins and conference wins. I may post some of our program strength, roster strength, and and recruiting scores later.
Note that A&M projects at #7, and is forecast at #2 in the SEC West. We move from slightly behind LSU in the pure team quality to higher expected wins becomes of our favorable schedule. Auburn comes in at 21 because of their heavy losses on offense and is the lowest rated team in the SEC West. I believe this is because most people highly underestimate the negative impact by losing so many receivers and their qb at the same time.
The high ranking of all the teams in the West translates into something that might surprise some; all teams in the SEC West have expected wins of less than 10. The reason is the brutal gauntlet of the West. Second, remember that there is uncertainty in these rankings. The actual success of A&M this year will depend on many factors, including close wins, turnovers, injuries, and the like. I am very confident that we will have a good year this season.
A couple of other teams of interest:
Arizona State #26 (we should win by a little more than 10)
Texas #51—should be a long year in Austin (expected wins: 4.6)
TCU #3 should handily beat Minnesota today (favored by 15.2 - 93.6% chance of winning)
-- If Doctson is out that is a big loss for TCU and their odds of winning would drop by around 10 percent.
I apologize for the messiness. I don't know how to create a clean table or I might include other data.
Wins listed below excluded bowl/playoff games. We use a home field advantage of 3 points:
Rank Team Conf. Exp. Wins Conf. Wins
1 Ohio State Big Ten 11.7 7.8
2 Alabama SEC 9.7 5.8
3 TCU Big 12 10.7 7.7
4 Georgia SEC 10.2 6.4
5 Oregon Pac-12 10.9 8.2
6 LSU SEC 9.2 5.2
7 Texas A&M SEC 9.3 5.5
8 Baylor Big 12 10.2 7.2
9 Ole Miss SEC 8.4 4.5
10 UCLA Pac-12 10.1 7.2
11 Oklahoma Big 12 9.0 6.6
12 Arkansas SEC 7.6 3.8
13 Tennessee SEC 8.5 4.9
14 Notre Dame Ind. 9.8 0.0
15 Clemson ACC 9.9 6.7
16 Florida SEC 8.3 4.6
17 Mississippi State SEC 7.5 3.7
18 Michigan State Big Ten 8.8 5.7
19 Florida State ACC 9.1 5.8
20 USC Pac-12 7.6 5.5
21 Auburn SEC 6.2 2.5
22 Nebraska Big Ten 9.1 5.7
23 Georgia Tech ACC 7.4 5.0
24 Arizona Pac-12 8.5 5.7
25 Wisconsin Big Ten 8.8 5.7
SEC West Exp. Wins Conf. Wins Div. Rank .
2 Alabama 9.7 5.8 1
7 Texas A&M 9.3 5.5 2
6 LSU 9.2 5.2 3
9 Ole Miss 8.4 4.5 4
12 Arkansas 7.6 3.8 5
17 Mississippi State 7.5 3.7 6
21 Auburn 6.2 2.5 7
Our Schedule and the probability of winning each game:
Opponent location Margin Probability of Winning
Arizona State Neutral 10.6 0.803
Ball State Home 32.1 1.000
Nevada Home 29.1 1.000
Arkansas Neutral 4.0 0.668
Mississippi State Home 8.8 0.784
Alabama Home -1.5 0.463
Ole Miss Away -0.3 0.500
South Carolina Home 14.9 0.933
Auburn Home 11.8 0.823
Western Caro. Home 1.000
Vanderbilt Away 15.9 0.938
LSU Away -3.7 0.342
Last edited:
