After looking through dozens of wikipedia season summaries one trend that jumps out is lots of early home games for teams like Michigan. 2016 is an extreme example: They didn’t play a road game until week 6. This year they just played their first road game in week 5. Why? And is it just Michigan? Since I can answer the second question with code let’s do that one first.
This recently committed home_games.py is a bare bones python executable using the cfbd API that takes advantage of the /games endpoint to roll up all results from a year into custom objects that aggregate the home/road/neutral game counts for each team. The widest net we can cast on /games is one whole year, so we do that and then there’s a big loop around it that takes a range of years. What do the last 10 years look like?
2011-2020 Team, home, road, neutral 1. Tennessee 70, 43, 5 2. Auburn 70, 44, 6 3. Alabama 68, 42, 14 4. Arkansas 67, 42, 9 5. Kentucky 70, 46, 2 6. LSU 67, 43, 8 7. Ohio State 67, 44, 8 8. Texas A&M 66, 43, 8 9. Penn State 68, 46, 4 10. Ole Miss 68, 46, 3 11. Clemson 69, 48, 8 12. Florida 62, 41, 15 13. Wisconsin 65, 45, 10 14. Purdue 66, 46, 2 15. Michigan 66, 46, 2 16. Notre Dame 63, 43, 13 17. Virginia 69, 49, 1 18. South Carolina 67, 47, 4 19. Georgia 63, 43, 16 20. Michigan State 67, 48, 3
Michigan is indeed on the list, but they’re all the way down at number 15. That scrappy underdog SEC is all over the top 10. Good for them doing whatever it takes to get an advantage with the big boys. Tennessee, at the top of the list, has averaged seven home games per year over the last decade with between four and five true road games.
What about past decades?
2001-2010 1. Auburn 73, 45, 1 2. Florida 71, 43, 6 3. LSU 74, 46, 1 4. Alabama 72, 45, 4 5. Nebraska 72, 48, 2 6. Penn State 70, 46, 1 7. Ohio State 71, 47, 0 8. Hawai'i 75, 51, 0 9. Arkansas 69, 46, 4 10. Michigan 70, 47, 0 11. Tennessee 71, 49, 0 12. Kansas 68, 46, 3 13. South Carolina 69, 48, 1 14. Texas A&M 68, 47, 2 15. Virginia Tech 69, 49, 5 16. NC State 69, 49, 0 17. Kansas State 69, 49, 1 18. Oklahoma 69, 50, 6 19. Kentucky 68, 49, 0 20. Michigan State 68, 49, 1 1991-2000 1. Hawai'i 78, 42, 0 2. Florida 70, 40, 0 3. Army 69, 41, 0 4. Nebraska 68, 44, 0 5. Auburn 66, 44, 0 6. Alabama 66, 44, 0 7. LSU 66, 44, 0 8. Florida State 67, 47, 0 9. Tennessee 65, 45, 0 10. Arkansas 64, 46, 0 11. Ohio State 66, 48, 0 12. Kansas State 64, 47, 0 13. Notre Dame 65, 48, 0 14. Illinois 63, 47, 0 15. Clemson 63, 47, 0 16. Michigan 64, 48, 0 17. Penn State 64, 50, 0 18. Georgia 62, 48, 0 19. Pittsburgh 62, 49, 0 20. North Carolina 62, 49, 0 1981-1990 1. Hawai'i 83, 33, 0 2. Army 72, 38, 0 3. South Carolina 72, 39, 0 4. Auburn 72, 39, 0 5. Arizona State 71, 39, 0 6. Georgia 69, 41, 0 7. Tennessee 69, 43, 0 8. Arizona 68, 42, 0 9. Kansas 67, 44, 0 10. Arkansas 66, 45, 0 11. Pittsburgh 65, 45, 0 12. Wisconsin 65, 46, 0 13. Nebraska 66, 47, 0 14. Penn State 65, 46, 0 15. LSU 64, 46, 0 16. Virginia Tech 64, 46, 0 17. Georgia Tech 64, 46, 0 18. Michigan 64, 47, 0 19. Ohio State 64, 47, 0 20. NC State 63, 47, 0
The 00s look very similar to the recent past, then things get weird. Hawaii I can understand (teams get a trip to Hawaii, Hawaii doesn’t bleed as much on airfare) but what’s up with Army?
Back to the recent past, there’s an obvious explanation for the top of this list: an arbitrage of money and prestige. All the programs on the top of the list have long-established public university programs with massively profitable home game infrastructure. They can guarantee a payout to a smaller school that exceeds the smaller school’s home gate and still retain enough from their own home gate to make it worth everyone’s while.
Auburn and Alabama are the only big sports programs fans in that state have to spend their time and money on. Arkansas is alone in its state. Kentucky competes with Louisville but no pro teams. LSU has one pro franchise in the state to compete with. Tennessee plays in the fifth-largest stadium in the United States. (In a state with one pro football team and no other big college team. Sorry Vandy.)
It’s interesting to look at wikipedia’s list of the biggest stadiums in the world. Of the top 10, 8 are US college football venues. Of those 8, 7 are in the recent unbalanced top-20 list. (The only one not on both lists is Texas, which has a line of 59, 45, 14 good for 38th place.)
For these programs with strong attendance and bottomless pools of fans to draw in it makes sense to pay for that extra game. Of course conferences schedules have to be balanced (uh I should check that) so these lopsided home stretches are more likely to happen at the beginning of the year, before conference play really ramps. Thus the observed phenomenon of Michigan’s home game September.
This does present an interesting betting scenario. More than once in the last few years an SEC team has grown fat on steady home and “neutral” games at the start of the year, risen in the rankings and the public’s enthusiasm, only to be rudely awakened in their first true road game. This past week we saw Arkansas fit this pattern. Their game at Georgia opened as ARK+18, the public bought the hype and bet it down to ARK+16.5 but the final was Georgia 37-0.
This should be something I can backtest with a little more code.