I've heard from a number of people that a two week layoff between the conference championship games and the super bowl usually leads to a blowout in the super bowl. I decided to go back over every super bowl and test this theory.
I found the line on every super bowl and compared that with the amount of time off before the super bowl. Only seven of the 41 super bowls have had a one week layoff, so there isn't very much to work with. Seven games doesn't come close to a big enough sample size for us to make any definite conclusions. Take everything from here on out with a grain of salt.
The first thing that jumps out at me about the seven one week games is that five of them were upsets. In the 34 two week games there were only four upsets. If we're looking at just recent history, there has been three one week super bowls in the last ten years and two of them were upsets (NE in 2001, TAM in 2002) while only one upset (DEN in 1997) in seven two week games. It does appear that a one week layoff is more conducive to a super bowl upset.
But do one week layoffs mean closer games? The average margin of victory in the seven one week super bowls is 10.1 points. Is that closer than normal? I didn't want to compare it to the average margin of victory of the regular season because that would include many games where a very bad team was facing a very good team. Theoretically, a super bowl should always be very good teams of somewhat equal strength. Instead, I calculated the average margin of victory of every playoff game in the super bowl era, which is 377 games. The average margin of victory in these games is 13.7. Oddly enough, the average margin of victory during regular season games since 1980 is just 11.5. It would appear that one week super bowls have produced closer games than a normal playoff game, but really aren't much different than a regular season game. I do find it interesting that playoff games, and super bowls in particular, have a larger margin of victory than regular season games, but we'll need to leave that for another blog. I can only assume that the better team is usually playing at home and is more focused in the playoffs.
Now let's compare this to the two week super bowls. The average margin of victory in two week super bowls is 16.4, well over the one week margin of victory and more than an average playoff game. We don't have much of a sample size to look at, but one week super bowls in general have been closer games than the two week variety.
The early line on this year's super bowl is 12 points so let's look at previous super bowls with double digit lines. There have been thirteen super bowls with double digit favorites. In those thirteen games, four of them resulted in upset wins for the underdog. Two of those came before the NFL/AFL merger, where not only did the general public undervalue the AFL, but the leagues did not play each other during the regular season making it very hard to gauge the better team. Still, even if we throw out those two games we get two upsets in eleven games. Believe it or not, those two upsets are the last two times we had a double digit line, NE in 2001 and DEN in 1997.
Let's look at how well the favorite has covered in the super bowl. The super bowl favorite is 25-14-2 against the spread. Recent history hasn't been so kind to the favorites, though. The super bowl favorite is just 4-5-1 in the last ten years.
None of this really means anything for this year's super bowl. You just can't conclude anything based on just 41 games over 41 years. The average margin of victory over the first 31 super bowls was 16.9 points, but it's been just 10.5 in the last ten seasons. If you play recent trends then a lot of the data points to a closer super bowl this season, regardless of how long the layoff.
Posted by Herb at 1/22/2008 1:58:00 PM