By Andrew Puopolo
This past weekend, Duke and Michigan were somewhat surprising winners of the ACC and Big 10 tournaments as 8 and 5 seeds, respectively. This made me wonder, are teams that win their conference tournaments over or under seeded in the NCAA Tournament? Does their momentum continue riding through March Madness, or do they regress back to the mean? Many will cite UConn winning 5 games in a row in the Big East Tournament as a 9 seed in 2011 and then winning March Madness as a 3, but where does that fit into the overall scheme of conference tournament winners?
The conferences used for this analysis were the Power 5 + Big East + American + Atlantic 10. The conferences were chosen because they consistently have earned more than one bid into the NCAA tournament. To test to see if there was a difference compared to seed, I took every team that won their conference tournament seeded 3rd or lower since 1997 (49 teams). I then used their expected number of games won in the tournament based on seed and compared it to the number of games actually won. From this, I ran a two sample t test to test to see if the means of the two samples were equal.
With a confidence interval of (-.14, .53), we cannot conclude that the means are different at a 95 percent confidence level and we fail to reject our null hypothesis that they were the same. However, it is possible that the paired test does not take outliers (like 2011 UConn) very well and that leads to skewed data. So, I decided to use a sign test to see if that would leave different results. Our results from the sign test were
Again, with a confidence interval of (-.39, .34) we find that we cannot reject the null hypothesis at a 95 percent confidence level. We conclude that surprise conference tournament winners perform no better or worse than their assigned seed. Therefore, we expect Duke to perform more or less like a 2 seed should and Michigan to perform around where a 7 seed should perform.