By Daniel Adler
The Raiders lost Super Bowl XXXVII after winning 11 games during the regular season. The next year, they won 5 games, setting in motion events that led to the return of Art Shell, the Lane Kiffin experiment, and the (ongoing) JaMarcus Russell era. Is there evidence that losing the Super Bowl can be devastating for a franchise?
USA Today recently (I need a straw man, okay?) ran an article sure to discourage Colts fans. They examine the 13 Super Bowl losers from 1993-2006 and determine that only 6/13 made the playoffs in the subsequent season and none advanced past the divisional round. This all sounds pretty dire, but what does it mean for the Colts? Is there a curse on the Super Bowl loser? Let’s examine the numbers.
Looking at the winners and losers of the first Super Bowls, it appears the both winners and losers make the playoffs in the following season at a similar rate.
|Playoffs Next Year||30/43=70%||28/43=65%|
Two more Super Bowl winners have returned to the playoffs than losers, but the difference is not statistically significant.
Furthermore, when we consider how the teams perform in the regular season compared to their previous season, we actually see that the losers more frequently improve.
|Record Next Season||Winners||Losers|
Of course, part of the reason for this fact is that Super Bowl winners have better regular season records on average. There is lots of reversion to the mean for both winners and losers.
Looking at the years 1970-2008, we see that the Super Bowl winners have had stronger regular season records than the losers and the gap is of equal magnitude in the next season.
|Regular Season Win Percentage (wins per 16 game season)*||Winners||Losers|
|Super Bowl Year||.807 (12.9 wins)||.748 (12.0 wins)|
|Next Season||.675 (10.8 wins)||.617 (9.9 wins)|
|Change||-.132 (-2.1 wins)||-.131 (-2.1 wins)|
*My data started in 1970 and did not include this season’s results.
So both teams win 2.1 fewer games on average, however, the winners are .9 better to start. It appears the winning or losing the Super Bowl does not predict next year’s success once we control for wins in the previous year.
A slightly more rigorous analysis incorporating multiple regression provides similar results. Once we control for previous year winning percentage or better yet, scoring ratio, winning/losing the Super Bowl has no significant predictive power for the next regular season winning percentage (.03, p=.31 for winner dummy; .01 p=.69 for loser dummy; i.e. not significant).
However, my straw man may have a bit of a point. Since 1993, 8/17 losers have made the playoffs while 12/17 winners have returned to the playoffs. Additionally, 14 of the losers got worse and only one improved. 11 winners won fewer games after winning the Super Bowl and three improved.
Still, the Colts should not worry about a Raiders-like tailspin. There are notable stories of post-Super Bowl loss struggles, but losing the Super Bowl has not negatively impacted a team’s chance of success in the subsequent regular season in a statistically significant way (5% level…or even 10%). As for the playoffs…we’ll look at that next time.
Thanks to Dan Yamins for his help with data collection. Also, thanks to Cade Massey and Richard Thaler for (tacitly) allowing me to use the title of their fantastic paper, The Loser’s Curse, which deals with the NFL draft. Definitely worth a read.