Ever since Alabama defeated LSU 21-0 in the BCS Championship game we have all heard the mantra that the rematch put LSU at a considerable psychological disadvantage. The idea is that after winning the first game, LSU felt like it had nothing to prove and therefore lacked motivation. Whereas the Alabama team was highly motivated by the loss to LSU. Of course it's easy play armchair psychologist, and one can easily come up with all sorts of theories on why one team or another is at an advantage or disadvantage. I could argue that the winning team has an advantage because they have confidence in their ability to win. But the bottom line is that when the ball gets snapped and it's man vs man, all of that junk goes out the window.
Anyway, it occurred to me that the Alabama/LSU rematch was far from the first rematch in a championship game in college football. So I wanted to look back at other games to see how the original winner and loser did. For now, I'm going to ignore bowl games because the data is a lot harder to assimilate. I'll just look at conference championship games which often end up being rematches.
SEC Championship Game-
There have been six rematches in the history of the SEC championship game. Years in which the same team won both games are bolded. Years in which the original game was decided by 8 points or less are marked with an *.
1999: RS- Alabama 40, Florida 39 (overtime). CG- Alabama 34, Florida 7 *
2000: RS- Florida 38, Auburn 7. CG- Florida 28, Auburn 6
2001: RS- Tennessee 26, LSU 18. CG- LSU 31, Tennessee 20 *
2003: RS- LSU 17, UGA 10. CG- LSU 34, UGA 13 *
2004: RS- Auburn 34, Tennessee 10. CG- Auburn 38, Tennessee 28
2010: RS- Auburn 35, USC 27 CG- Auburn 56, USC 17 *
ACC Championship- The ACC championship has featured four rematches.
2007: RS- BC 14, VT 10. CG- VT 30, BC 16 *
2008: RS- BC 28, VT 13. CG- VT 30, BC 12
2009: RS- GT 30, Clemson 27. CG- GT 39, Clemson 34 *
2011: RS- Clemson 23, VT 3. CG- Clemson 38, VT 10
Big 10 Championship Game-
2011: RS- MSU 37, Wis 31. CG- Wis 42, MSU 39 *
So what does all of this data, limited as it is, show? In 7 out of 11 rematches (64%), the rematch ended with the same team winning. You could argue that the data is skewed because if one team is significantly better than the other, then they are likely to win the rematch regardless of any psychological advantage the other team might gain. However if you only look at examples where the first game was close (take out games where the margin of victory was greater than 8 points), the original winner is still 4-3 (57%) in the rematches. And in 3 of those where the orignal game was very close, the winner ended up blowing out their opponent in the rematch.
For another way to look at the data, I calculated the change in margin of victory for the winner of the first game compared to their performance in the second game. So a positive value correlates to the original winner winning by more points in the rematch. A negative value indicates that the rematch was closer than the first game or that the original loser won the rematch. The average change in margin of victory in the 11 championship game rematches was -1.9. This does indicate that the losing team performs better in a rematch. Slightly. However, they are still more likely to lose the rematch (36% win rate).
Anyway, in the end I think that the advantage for a losing team in a rematch is largely overblown. It's mostly perpetuated by confirmation bias. In other words, it's a weak excuse for getting your ass whooped.