College Football Computer Ratings after Week 12
Rating all 134 FBS teams and evaluating the college football playoff selection committee's rankings
I've posted ratings of NFL teams for several weeks, and my college football ratings are similar but not quite identical. The NFL and college football ratings are generated from two different sets of source code. The actual process is very similar, but the college football ratings include some tweaks that I'll introduce into my NFL ratings after the season. Before posting the ratings, I want to provide some insight into how this process actually works. The general process for my college football predictive ratings that I describe here is also what I use for the NFL ratings, just with 50 trials to calculate the ratings instead of 30, and with a few differences in details beyond what I discuss in this article.
The ratings are probably best viewed on a desktop computer or a tablet, where the display is sufficiently wide for tables to render properly. I'm working on some other ways to display these ratings, but for the moment, this works fairly well on desktop displays though not necessarily on mobile devices.
Predictive Ratings
The predictive ratings my system calculates are forward-looking and only consider margin of victory. I'll explain a bit later why this is better than looking at wins and losses, but I want to first explain where these ratings come from. The goal of the rating system is to calculate a rating for each team and an overall home field advantage that would best predict the margin of victory for the games that have been played so far in the season. The margin of victory is predicted by taking the home team's rating, adding the home field adjustment if it's a home game instead of at a neutral site, and then subtracting the visiting team's rating. In this example, Team A has a rating of 65, Team B has a rating of 55, and the home advantage is 2.5 points. If Team A is the home team, then their expected margin of victory is 65 + 2.5 - 55, or 12.5 points. If Team B is the home team, then their expected margin of victory is 55 + 2.5 - 65, which is -7.5 points, meaning that Team B is expected to lose by 7.5 points. At a neutral site, Team A's expected margin of victory is just 65 - 55, or to win by 10 points.
The rating system makes these predictions for every game that's been played, keeping track of how accurate its predictions are for each team. If a team consistently plays worse than the ratings predict it should, it's an indication that the team is overrated and that its rating should probably be lowered a bit. However, a team is probably underrated if its margin of victory is typically better than the predictions, and it's a sign that the rating likely needs to be adjusted upward. The same approach is also used to determine if home teams systematically overperform or underperform the predictions. The system also adds in a small random adjustment because this seems to be helpful in trying to get closer to optimal ratings for the teams.
When the rating process begins, the first guess for each team's rating and the overall home advantage are all set to zero. The ratings are then nudged up or down as I described. It's no surprise that when all the ratings are zero, the predictions aren't very good. After a several thousand passes of adjusting the ratings, it becomes very difficult to find any new adjustments that will make the ratings better. At this point, the rating process stops, keeping the set of ratings that best predicted the outcome of the games that have been played for far during the season.
It's time to ask an important question: because there's a bit of randomness each time the ratings are adjusted, does this mean using different random numbers will produce different ratings? The answer is definitely yes. The differences are often quite small, but sometimes the ratings can vary by a couple of points in either direction. Because I don't want random chance to skew the ratings, I actually repeat the entire process 30 separate times, each time resetting the initial ratings to zero, then using different random adjustments. This means that I get 30 different sets of predictive ratings. The predictive rating that's displayed is the median of these 30 different trials. There's still a bit of randomness to the ratings, but its impact is small enough that I'm not too concerned.
Strength of Record Ratings
It's time to ask another important question: why are points better at predicting future performance than wins and losses? From the perspective of the predictive ratings, winning a football game by one point doesn't look that much better than losing a game by a single point. But in the standings, that two point swing makes all the difference. Shouldn't we give more credit to teams that find ways to win close games?
When I introduced my NFL ratings, I linked to a YouTube video explaining that there's a lot of luck in the outcome of NFL games. I have no reason to believe this is any different for college football games. This luck doesn't necessarily even out over the course of 12 or even 17 games, meaning that the standings aren't necessarily the best indicator of how good a team is. If a team wins most of their games by a large margin, it's a sign that they really are better than their competition. Likewise, there's probably little doubt that a team that routinely gets blown out is almost certainly not as good as their opponents. But what about a team that wins a lot of close games, or one that loses a lot of close games? Is a team that's 9-1 in games decided by a touchdown or less clearly better than their opponents? More likely, they're probably fairly similar to their competition, perhaps slightly better. Wins and losses aren't necessarily indicative of how good a team really is, which is why margin of victory is a better predictor of future performance.
However, when we determine which teams qualify for conference championship games or postseason play, this is based on the standings. Even if wins and losses are not the best predictor of future performance, we use them to determine which teams are most deserving of playing in postseason tournaments. For a league like the NFL, schedules are balanced enough across the league that wins and losses are sufficient to decide the most deserving teams. However, college teams play schedules of vastly different strengths, so we need a better way to determine which teams' records are most impressive. That's where strength of record ratings come in.
My strength of record ratings are based on a team's actual winning percentage, then subtracting the expected winning percentage if some "other team" played the exact same schedule. If a team's actual winning percentage is better than the expected winning percentage, then their record is strong, and they should be rated highly. If not, their record is weaker, and they will be rated lower. ESPN’s FPI uses a bit of a different approach to calculating strength of record, but it’s a similar concept.
It's time to ask another question: what hypothetical "other team" is used for the expected winning percentage? Although I only show ratings for FBS teams in this article, my system actually calculates ratings for nearly all FBS, FCS, Division II, and Division III teams, as well as some non-NCAA opponents. The average predictive rating across all divisions is almost exactly zero. The mean predictive rating for all FBS teams is 45.42. It turns out that the strength of record ratings and the ranking of the teams can be quite different depending on if the "other team" has a zero or 45.42 rating. I actually calculate the strength of record using both of these benchmark ratings, then for six other benchmark ratings as well, and average them all together. They are then adjusted statistically to be on the same scale as the predictive ratings.
Overall Ratings and Schedule Strength
In some cases, there's a big difference between the predictive and strength of record ratings. The 4-6 Kansas Jayhawks are ranked #72 on strength of record but #29 on predictive ratings. Both the polls and the college football playoff selection committee take into account strength of record and margin of victory in how they rank teams. This was very obvious when the selection committee controversially picked one-loss Alabama over undefeated Florida State last year. Florida State was an unbeaten champion of a Power Five conference, but Alabama had stronger predictive ratings and was the champion of an overall stronger conference. The overall rating also incorporates the quality of a team's record and expected future performance by averaging the predictive and strength of record ratings, so it could also be an approximation for how the selection committee could rank teams.
None of these ratings explicitly account for schedule strength. That's another topic, one with a lot of detail that's worthy of its own article. A lot of schedule strength in college football depends on which conference a team is in, and I'll discuss which conference is best in that article. Spoiler: it's the SEC by a mile, but the second best conference is a much more interesting discussion. For the purpose of these ratings, however, schedule strength is simply the average of the predictive ratings for a team's opponents to date, adding in an adjustment for home field advantage.
The Ratings
You've probably had enough of me explaining where these ratings come from, and you'd like me to just post the ratings already. The "Rating" column is the overall rating, which is the average of the predictive and strength of record ratings. The "Predictive" column is the predictive rating, with the rank in parentheses. Strength of record is in the "Record" column, with the rank also in parentheses. The "Schedule" column works the same way, but for schedule strength. The home advantage in this rating cycle is 2.52 points. To calculate the expected of margin of victory in a game, take the home team's rating, add the home advantage, then subtract the road team's rating. The predictive rating is probably going to be the best estimate of a team's future performance.
After all that discussion, here are the ratings for all 134 FBS teams for games played through November 17, 2024:
Overall Ratings
Home advantage: 2.52 points
Rank Rating Team Predictive Record Schedule
1 72.29 Alabama 76.29 (1) 68.29 (7) 53.90 (7)
2 71.97 Texas 73.74 (3) 70.19 (3) 48.91 (33)
3 71.35 Georgia 71.31 (5) 71.39 (2) 59.04 (1)
4 69.89 Ohio State 72.69 (4) 67.10 (10) 45.04 (58)
5 69.23 Oregon 64.60 (12) 73.87 (1) 46.54 (51)
6 69.00 Notre Dame 71.12 (6) 66.88 (11) 44.79 (60)
7 68.98 Ole Miss 74.15 (2) 63.82 (17) 47.12 (48)
8 68.20 Indiana 67.67 (8) 68.73 (5) 39.09 (87)
9 66.92 Penn State 65.21 (11) 68.62 (6) 46.25 (53)
10 66.82 Miami 66.06 (10) 67.58 (8) 44.61 (62)
11 66.29 Tennessee 67.83 (7) 64.75 (14) 48.95 (32)
12 65.42 South Carolina 67.06 (9) 63.77 (18) 55.55 (4)
13 64.73 BYU 59.87 (20) 69.59 (4) 47.56 (45)
14 64.56 SMU 61.88 (16) 67.23 (9) 44.07 (67)
15 64.12 Colorado 62.15 (14) 66.08 (12) 50.50 (20)
16 62.93 Clemson 62.10 (15) 63.77 (19) 48.14 (41)
17 62.61 Texas A&M 60.50 (19) 64.72 (15) 48.60 (38)
18 62.52 Iowa State 61.14 (17) 63.89 (16) 47.74 (42)
19 61.49 Arizona State 57.56 (28) 65.41 (13) 49.61 (29)
20 60.90 Boise State 58.80 (24) 62.99 (20) 39.79 (82)
Rank Rating Team Predictive Record Schedule
21 60.35 Tulane 62.62 (13) 58.09 (24) 39.18 (85)
22 59.96 Kansas State 59.04 (22) 60.89 (21) 51.30 (14)
23 58.83 LSU 59.44 (21) 58.22 (23) 55.37 (6)
24 57.65 Louisville 60.53 (18) 54.76 (32) 50.63 (18)
25 57.03 Army 53.98 (40) 60.08 (22) 28.76 (133)
26 56.88 Missouri 55.67 (34) 58.09 (25) 47.42 (46)
27 55.87 Florida 58.31 (26) 53.44 (36) 56.67 (2)
28 55.82 Iowa 59.03 (23) 52.62 (42) 48.26 (40)
29 55.49 Pittsburgh 54.42 (36) 56.57 (27) 44.51 (63)
30 55.30 UNLV 54.19 (39) 56.42 (28) 37.70 (95)
31 55.08 Vanderbilt 55.74 (33) 54.42 (33) 49.67 (28)
32 54.84 Baylor 56.33 (32) 53.35 (37) 48.71 (35)
33 53.91 USC 58.62 (25) 49.20 (57) 50.58 (19)
34 53.74 Oklahoma 56.48 (31) 51.00 (45) 53.54 (8)
35 53.61 Virginia Tech 57.61 (27) 49.60 (52) 51.30 (15)
36 53.57 Syracuse 49.99 (55) 57.16 (26) 46.16 (54)
37 53.34 Minnesota 55.26 (35) 51.43 (44) 46.08 (55)
38 53.24 Georgia Tech 53.20 (43) 53.27 (38) 47.68 (43)
39 52.91 Illinois 50.18 (54) 55.64 (30) 43.87 (69)
40 52.43 Washington State 49.00 (59) 55.86 (29) 37.86 (94)
Rank Rating Team Predictive Record Schedule
41 52.16 TCU 52.01 (45) 52.32 (43) 46.71 (49)
42 52.10 James Madison 51.13 (47) 53.07 (39) 33.72 (122)
43 51.91 Arkansas 54.35 (37) 49.48 (55) 49.75 (27)
44 51.77 Texas Tech 49.59 (57) 53.95 (35) 49.42 (30)
45 51.76 Duke 48.34 (60) 55.18 (31) 43.98 (68)
46 51.57 Wisconsin 53.83 (41) 49.32 (56) 51.32 (13)
47 51.11 Cincinnati 53.39 (42) 48.84 (58) 50.05 (25)
48 50.86 Kansas 56.98 (29) 44.74 (72) 51.73 (12)
49 50.47 Washington 50.90 (49) 50.04 (49) 48.65 (37)
50 50.32 North Carolina 51.13 (48) 49.51 (53) 43.86 (70)
51 50.06 UCF 56.76 (30) 43.36 (78) 50.21 (24)
52 50.00 Louisiana 46.02 (66) 53.98 (34) 34.31 (115)
53 49.98 West Virginia 50.33 (52) 49.64 (51) 50.45 (21)
54 49.63 Michigan 49.29 (58) 49.97 (50) 52.12 (10)
55 49.18 Rutgers 47.54 (61) 50.82 (46) 44.27 (64)
56 49.03 Memphis 45.15 (70) 52.91 (41) 32.53 (124)
57 48.61 Boston College 50.47 (50) 46.75 (65) 47.19 (47)
58 48.36 Navy 46.26 (65) 50.47 (47) 37.64 (97)
59 48.25 Kentucky 52.87 (44) 43.62 (77) 50.36 (23)
60 47.91 Nebraska 49.84 (56) 45.98 (66) 46.58 (50)
Rank Rating Team Predictive Record Schedule
61 47.76 California 51.72 (46) 43.80 (76) 44.08 (66)
62 47.73 Marshall 45.35 (69) 50.12 (48) 38.09 (93)
63 47.68 Virginia 47.07 (63) 48.29 (60) 49.89 (26)
64 47.42 Auburn 54.27 (38) 40.58 (90) 46.28 (52)
65 47.20 UCLA 47.07 (62) 47.32 (62) 55.51 (5)
66 46.51 Texas State 50.36 (51) 42.66 (80) 34.86 (111)
67 46.16 Georgia Southern 42.83 (77) 49.49 (54) 44.62 (61)
68 45.75 Utah 50.30 (53) 41.20 (89) 47.64 (44)
69 45.60 Sam Houston 38.16 (92) 53.04 (40) 33.92 (119)
70 45.39 Toledo 43.18 (74) 47.60 (61) 34.04 (118)
71 44.32 Western Kentucky 41.41 (84) 47.22 (63) 33.77 (120)
72 44.23 Houston 43.64 (73) 44.81 (70) 52.09 (11)
73 43.70 Maryland 45.59 (68) 41.81 (87) 48.69 (36)
74 43.67 UConn 41.59 (83) 45.74 (67) 31.51 (127)
75 43.58 Bowling Green 43.07 (76) 44.09 (74) 35.78 (103)
76 43.44 Ohio 40.08 (86) 46.79 (64) 32.12 (126)
77 43.36 Northwestern 44.37 (72) 42.36 (83) 48.90 (34)
78 43.35 Miami (OH) 42.60 (80) 44.10 (73) 36.77 (99)
79 43.04 NC State 42.22 (81) 43.85 (75) 44.21 (65)
80 42.78 Michigan State 42.73 (79) 42.82 (79) 48.48 (39)
Rank Rating Team Predictive Record Schedule
81 42.52 Colorado State 36.33 (98) 48.71 (59) 35.37 (106)
82 42.50 Arizona 42.80 (78) 42.20 (86) 49.00 (31)
83 42.28 Oklahoma State 46.35 (64) 38.21 (98) 50.76 (16)
84 42.18 Jacksonville State 39.57 (88) 44.78 (71) 30.30 (131)
85 42.05 Northern Illinois 41.60 (82) 42.51 (81) 34.69 (114)
86 41.87 Old Dominion 44.82 (71) 38.91 (96) 45.24 (57)
87 40.64 South Alabama 43.17 (75) 38.12 (99) 34.86 (110)
88 40.64 South Florida 39.94 (87) 41.33 (88) 39.93 (80)
89 40.06 Wake Forest 40.32 (85) 39.80 (93) 44.83 (59)
90 39.63 San José State 36.78 (96) 42.48 (82) 34.71 (113)
91 39.41 North Texas 38.73 (90) 40.10 (91) 38.84 (89)
92 39.36 Mississippi State 45.61 (67) 33.10 (107) 52.30 (9)
93 39.13 Coastal Carolina 38.28 (91) 39.99 (92) 39.16 (86)
94 38.52 UL Monroe 34.73 (100) 42.32 (84) 41.93 (73)
95 38.51 Liberty 31.90 (108) 45.13 (69) 24.72 (134)
96 38.10 East Carolina 36.53 (97) 39.68 (94) 30.50 (130)
97 37.87 Stanford 37.74 (94) 38.01 (100) 50.44 (22)
98 37.82 Fresno State 38.05 (93) 37.59 (101) 35.65 (104)
99 37.49 Arkansas State 29.53 (118) 45.45 (68) 38.45 (91)
100 36.36 App State 33.71 (103) 39.01 (95) 42.04 (72)
Rank Rating Team Predictive Record Schedule
101 36.27 Buffalo 30.31 (113) 42.24 (85) 34.06 (117)
102 36.04 UTSA 34.54 (102) 37.53 (102) 35.20 (107)
103 35.55 Western Michigan 32.78 (106) 38.33 (97) 35.12 (108)
104 34.71 Florida State 39.05 (89) 30.37 (110) 55.86 (3)
105 34.08 New Mexico 32.68 (107) 35.49 (103) 36.77 (100)
106 33.56 Rice 37.11 (95) 30.01 (111) 39.32 (84)
107 33.32 Oregon State 32.87 (105) 33.77 (106) 38.32 (92)
108 32.55 Eastern Michigan 30.69 (111) 34.40 (104) 30.51 (129)
109 32.34 Nevada 36.16 (99) 28.52 (115) 41.02 (78)
110 32.00 Charlotte 30.13 (114) 33.87 (105) 45.47 (56)
111 31.86 Troy 34.68 (101) 29.05 (113) 38.98 (88)
112 30.75 Hawai'i 30.60 (112) 30.90 (108) 35.37 (105)
113 30.65 Utah State 31.45 (110) 29.85 (112) 39.84 (81)
114 28.85 San Diego State 29.97 (115) 27.74 (116) 36.86 (98)
115 28.84 Georgia State 31.84 (109) 25.83 (121) 42.60 (71)
116 28.79 Florida International 33.24 (104) 24.34 (125) 33.77 (121)
117 28.40 Air Force 29.61 (117) 27.18 (118) 36.77 (101)
118 28.40 Louisiana Tech 29.66 (116) 27.13 (119) 28.88 (132)
119 28.17 Purdue 29.44 (119) 26.89 (120) 50.69 (17)
120 28.05 Temple 25.57 (125) 30.53 (109) 41.12 (76)
Rank Rating Team Predictive Record Schedule
121 27.89 Ball State 26.80 (122) 28.97 (114) 39.34 (83)
122 26.76 UAB 28.45 (120) 25.07 (124) 40.23 (79)
123 26.22 Central Michigan 26.67 (123) 25.77 (122) 34.85 (112)
124 25.05 Wyoming 27.71 (121) 22.39 (127) 38.50 (90)
125 24.91 Akron 24.58 (126) 25.25 (123) 41.56 (74)
126 24.25 Middle Tennessee 21.19 (129) 27.31 (117) 37.65 (96)
127 23.23 Florida Atlantic 25.94 (124) 20.53 (129) 34.95 (109)
128 21.32 Massachusetts 22.50 (127) 20.14 (130) 33.54 (123)
129 21.01 Tulsa 18.07 (133) 23.95 (126) 31.22 (128)
130 20.02 UTEP 21.91 (128) 18.14 (132) 32.53 (125)
131 19.92 New Mexico State 18.87 (132) 20.98 (128) 36.71 (102)
132 19.24 Southern Miss 19.41 (130) 19.07 (131) 41.02 (77)
133 16.31 Kennesaw State 19.26 (131) 13.36 (134) 34.22 (116)
134 12.77 Kent State 11.46 (134) 14.07 (133) 41.21 (75) College Football Playoffs
The college football playoff committee's rankings came out just over 30 hours ago. Perhaps the obvious question is whether the committee got it right. Unfortunately, I can't answer that question in anything resembling an objective manner. Their rankings inherently include an element of subjectivity, trying to achieve some balance between which teams are the best teams and which are most deserving based on their records and schedules. Instead, I'm going to try to answer which teams might have the biggest complaints statistically about where they're ranked.
The overall ratings are my attempt to combine which teams are most deserving (strength of record rating) and which are best (predictive rating) in a single rating. However, if the ratings are purely based on which teams are most deserving based on schedule strength, wins, and losses, then strength of record is a better tool.
Georgia is ranked #3 in the overall ratings at 71.35. The playoff committee has them Georgia all the way down at #10, which is where Miami is ranked with an overall rating of 66.82. For Georgia to be ranked at #10 in the overall ratings, they would have to be ranked 4.53 points lower. This is the basic approach I'll use to calculate which teams I believe are most out of place in the committee's rankings. For teams that are in my top 25 but aren't ranked by the committee, I'll compare them to the #26 team in my ratings. To some extent, a team's objections depend on whether the committee is evaluating them only on strength of record, or if they're trying to also factor in which are the best teams. Therefore, I'll consider both the strength of record and overall ratings.
Strength of Record Rating Comparison
Here are the differences between the committee’s rankings and the strength of record ratings:
Rank Difference Team Actual Effective
1 -4.84 BYU 69.59 64.75
2 -4.52 Arizona State 65.41 60.89
3 -4.30 Georgia 71.39 67.10
4 -3.81 Iowa State 63.89 60.08
5 -3.74 Kansas State 60.89 57.16
6 -2.19 Colorado 66.08 63.89
7 -1.82 SMU 67.23 65.41
8 -1.07 LSU 58.22 57.16
9 0.00 Oregon 73.87 73.87
10 0.00 Texas 70.19 70.19
11 0.00 Indiana 68.73 68.73
12 0.00 Alabama 68.29 68.29
13 0.00 Miami 67.58 67.58
14 0.00 Texas A&M 64.72 64.72
15 0.00 South Carolina 63.77 63.77
16 0.05 Clemson 63.77 63.82
17 0.13 Missouri 58.09 58.22
18 0.97 Penn State 68.62 69.59
19 1.67 UNLV 56.42 58.09
20 1.74 Notre Dame 66.88 68.62
Rank Difference Team Actual Effective
21 2.13 Tennessee 64.75 66.88
22 2.44 Illinois 55.64 58.09
23 3.10 Boise State 62.99 66.08
24 3.41 Ole Miss 63.82 67.23
25 3.68 Army 60.08 63.77
26 4.30 Ohio State 67.10 71.39
27 4.90 Tulane 58.09 62.99 Overall Rating Comparison
And here are the differences between the committee’s rankings and the overall ratings:
Rank Difference Team Actual Effective
1 -4.53 Georgia 71.35 66.82
2 -3.31 Alabama 72.29 68.98
3 -3.09 Kansas State 59.96 56.88
4 -2.90 South Carolina 65.42 62.52
5 -2.55 Iowa State 62.52 59.96
6 -2.06 Ole Miss 68.98 66.92
7 -1.95 LSU 58.83 56.88
8 -1.19 Colorado 64.12 62.93
9 -1.14 Arizona State 61.49 60.35
10 -0.77 Louisville 57.65 56.88
11 -0.62 Texas 71.97 71.35
12 -0.32 Clemson 62.93 62.61
13 -0.18 BYU 64.73 64.56
14 0.00 Notre Dame 69.00 69.00
15 0.00 Tennessee 66.29 66.29
16 0.18 SMU 64.56 64.73
17 0.54 Tulane 60.35 60.90
18 1.03 Indiana 68.20 69.23
19 1.38 Miami 66.82 68.20
20 1.51 Texas A&M 62.61 64.12
Rank Difference Team Actual Effective
21 1.95 Missouri 56.88 58.83
22 2.07 Ohio State 69.89 71.97
23 2.34 UNLV 55.30 57.65
24 2.97 Penn State 66.92 69.89
25 3.06 Oregon 69.23 72.29
26 4.12 Illinois 52.91 57.03
27 4.46 Army 57.03 61.49
28 4.52 Boise State 60.90 65.42 Which Teams Were Ranked Most out of Place?
The "Actual" column is a how my system actually rated each team. The "Effective" column is the rating each team would need to have based on where the selection committee ranked them. The "Difference" column is the difference between the "Effective" and "Actual" columns, where a negative number means that the committee underrated a team compared to my ratings, and a positive number means the committee overrated the team compared to my ratings. Because the committee's objectives are a bit subjective and ambiguous, I'm going to focus on teams that were overrated according to both ratings or underrated according to both ratings.
Georgia has the strongest complaint about where they're ranked, and it's really not that close. The committee has them at #10 whereas my ratings have them at #2 and #3. Yes, they have two losses, but they've also played the toughest schedule according to my ratings. After Georgia, the four Big 12 schools of Arizona State, Iowa State, Colorado, and Kansas State have the strongest complaints. Although BYU's case for being underrated is weaker, they're still a bit underrated compared to the overall ratings, and they're the most undervalued team according to the strength of record ratings. Despite some grumbling about the committee underrating the SEC, it's actually the Big 12 that seems to be systematically underrated by the committee. By the way, when I said the SEC is the best conference this year, there's a strong statistical argument that the Big 12 is actually the second strongest conference when all the teams are considered.
However, LSU also has an argument that they've been underrated by the selection committee, in that they've been completely left out of the top 25. Their 6-4 record doesn't look as impressive as some other teams, but it's been compiled against the #6 ranked schedule. ESPN's FPI ratings (#19), the FPI strength of record (#20), and Jeff Sagarin's ratings (#18) all place LSU above any of my ratings, so it's possible that I'm still underrating them to an extent.
I don't like picking on Group of Five teams, but there's a statistical argument that Army, Boise State, Tulane, and UNLV are all overrated to some degree. Otherwise, the Big Ten trio of Ohio State, Penn State, and Illinois are all ranked higher by the committee than in my ratings. Despite the aforementioned grumbling about the committee disrespecting the SEC, it's a bit surprising to see Missouri still ranked at #23. Although their game against South Carolina was closer than predictive ratings would have indicated, Missouri found a way to lose the game at the end. This is a bit smaller of a difference than some of the other teams I mentioned, but Missouri is probably slightly overrated by the selection committee.
I'm in favor of the committee evaluating a team's resume instead of estimating a team's future performance, meaning that I prefer the strength of record ratings. With that in mind, here are my seeds for the playoff:
Playoff Seeds
#1 Oregon
#2 Georgia
#3 BYU
#4 Miami
#5 Texas
#6 Indiana
#7 Penn State
#8 Alabama
#9 SMU
#10 Ohio State
#11 Notre Dame
#12 Boise State
First Four Out
#13 Colorado
#14 Arizona State
#15 Tennessee
#16 Texas A&MWhat’s Next?
Now that I have my college football ratings working satisfactorily, I'll publish the next round of ratings before Tuesday's games. I also want to take a more detailed look at schedule strength and conference quality, which will go beyond what I've included here. I'll also be posting my NFL ratings and picks ahead of the Thursday night game. I'm planning to post some baseball-related content ahead of and during the winter meetings, which begin less than three weeks from now.
If you're interested in more content like this, please click the "Subscribe" button below and share this article on social media. I develop my own code to generate these ratings and predictions, of course relying on data from others and open source libraries, but none of my content is or will ever be created with generative AI. Your help in supporting more quality content like this is greatly appreciated. This article relies on data from CollegeFootballData.com, which contains a wealth of information and data about college football, makes it freely available, and shares data in formats that are easily read by many software tools.