Bowl Game Rematch Trends Point To Alabama Beating LSU In A Low Scoring Defensive Battle

December 8, 2011 - by David Hess

We’ve seen LSU play Alabama once already this season. In fact, they play every year, as you can see on our LSU-Alabama head-to-head results page.

Don’t worry, we’re not here to complain about the fact that the BCS Title Game is a rematch. We’re here to learn from it.

Bowl game rematches are relatively rare. We only have seven in our database, which goes back to 2003:

Date	Bowl	Score of First Game	Bowl Result	Same Winner?
1/1/04	Gator	Maryland 34-7 West Virginia	Maryland 41-7 West Virginia	Yes
1/1/04	Orange	Miami 22-14 Florida St	Miami 16-14 Florida St	Yes
12/22/07	Las Vegas	UCLA 27-17 BYU	BYU 17-16 UCLA	No
12/26/07	Motor City	Purdue 45-22 Central Michigan	Purdue 51-48 Central Michigan	Yes
12/31/08	Armed Forces	Air Force 31-28 Houston	Houston 34-28 Air Force	No
12/20/08	Eagle Bank	Navy 24-17 Wake Forest	Wake Forest 29-19 Navy	No
12/30/10	Holiday	Nebraska 56-21 Washington	Washington 19-7 Nebraska	No
1/9/11	BCS Title	LSU 9-6 Alabama	?	?

Well, there’s a surprise — the winner of the first game won only 3 of 7 rematches!

You stat geeks out there probably already know that people often place too much emphasis on head-to-head results when evaluating teams. The tendency is to assume that the winner on the field “proved” that they were the better team. Yet check out some of the turnarounds above. Last year Washington lost at home to Nebraska by 5 touchdowns, but won the bowl rematch 19-7.

OK, so the first take away is that we shouldn’t put too much stock in the fact that LSU won the first game. What can we learn by looking at the betting data?

This table shows the spread and totals lines for the initial games and the bowl rematches. It also shows the final scores, which team covered, and whether the score went Over or Under the total. In every case, the winner of the first matchup is “Team 1” and the spread is from their perspective. (So, a +2 spread in the bowl game means that the loser of the first game was favored by two in the bowl matchup.)

Team 1 = Winner of first game		BOWL GAME LINES & RESULTS						FIRST GAME LINES & RESULTS
Team 1	Team 2	Team 1 Spread	O/U	Pts 1	Pts 2	Cover	O/U	Team 1 Spread	O/U	Pts 1	Pts 2	Cover	O/U
Maryland	W Virginia	-4.5	50	41	7	1	U	-9.5	46.5	34	7	1	U
Miami	Florida St	+2	46	16	14	1	U	+6.5	49	22	14	1	U
UCLA	BYU	+6	46	16	17	1	U	-8.5	47.5	27	17	1	U
Purdue	C Michigan	-7.5	71.5	51	48	2	O	-21.5	67	45	22	1	P
Air Force	Houston	+4	66	28	34	2	U	+2.5	51.5	31	28	1	O
Navy	W Forest	+3	44	19	29	2	O	+17	51.5	24	17	1	U
Nebraska	Washington	-11.5	54.5	7	19	2	U	-3	51	56	21	1	O
LSU	Alabama	+1	40	?	?	?	?	+5	41	9	6	1	U

There are several interesting tidbits here:

The same team was favored in both games 6 of 7 times, including all three times that the underdog won in game one. This is an indication that the betting markets understand what we mentioned above — you have to judge the season as a whole, not based off the results of just one game, even if that one game is a head to head clash.
The totals line has been an average of 5 points too high in the rematches. In the first games, even though the Under went 4-2-1, the average score was extremely close to the average totals line: 52.1 actual points scored per game, versus 52.0 average totals line. In the rematches, the average totals line was 54.0, but the average points scored were only 49.4. Now, the sample size is minuscule, so don’t put much stock in it. But it at least seems plausible that the second time around, one of the defenses might figure out how to stop a key aspect of the opponent’s offense, lowering the scores more than expected.
The loser of game one has covered the bowl spread the last 4 times. Of course, the winner covered the last 3 before that. This is more than likely just noise, but it’s interesting nonetheless.
The winner covered the spread in 5 of 7 bowl rematches. Of course, that’s extremely likely this year, as the spread currently sits as Alabama -1. The only result that wouldn’t have the winner covering is Alabama winning by 1.
Whichever direction the totals line moves from the first game to the second, the score usually does as well (5 of 7 times). This trend will probably not hold true. The totals line has moved down from 41 to 40, but it’s unlikely the final score will be lower than 9-6.

Make of these what you will. Our main goal here was to get this data out in the wild, so you can use it as you see fit.

One thing remains clear — the BCS Title Game is expected to be a close, low scoring, defensive battle. Nothing in the above trends points us in any other direction.

By the way, if you’re going to compete in a Bowl Pick’em or are planning on wagering on some of the games, make sure to check out our 2011-2012 college football bowl games section, complete with bowl predictions, bowl betting picks, and bowl pick’em advice and picks.