I did a similar data analysis of kenpom a few weeks ago. While AdjEM is a great metric for determining champions and finalists, I’d imagine the numbers you’re using are the end of season metrics, which include massive swings from the NCAA tourney itself, given the elite competition you face and the incredible performances you need to have to win.A few other KenPom data points to add to the mix:
NCAA Champion.
- The NCAA tournament champion has been one of KenPom's top 3 teams in adjusted efficiency margin in 16 of the 19 tournaments held since 2002, and the season leader in adjEM has won 10 of the 19 tournaments.
- The top 3 teams in adjEM this season (to date) are Gonzaga (32.65), Arizona (27.45) and Kentucky (27.32). (AdjEM is the difference between a team’s offensive and defensive efficiency, and represents the number of points the team would be expected to outscore the average D-I team over 100 possessions).
- The average NCAA champion since 2002 had an adjEM of 30.45.
- Every champion since 2002 has had an adjEM above 28.2 except for 3 outliers (2014 UConn (22.13); 2003 Syracuse (23.38); and 2011 UConn (23.93).
- Fortuitous timing helps: Of the 21 years of date KenPom has published since 2002, 11 years had a single team in the field with an adjEM >32, 6 years had no teams with a 32 rating or better, and 4 years had more than one great team with a rating over 32.0 (including, sadly, 2005). The best years for a potential darkhorse champion (including all 3 of the outliers noted above) are years with no teams above 32 or a single team above 32 that is upset in the tournament.
Lowest Rated Final Four Team
- The average adjEM of the lowest rated Final 4 team each year since 2002 is 21.96, and in 15 of 19 years the adjEM of the lowest Final Four team was above 20.1. (The four outliers were 2010 VCU (13.49); 2018 Loyola (16.39); 2006 George Mason (18.16); and 2016 Syracuse (18.57).
Illini adjEM season to date is 19.91.
Looking at champions and their pre-tourney (solely regular season) AdjEM tells a much different story. Only 3 out of 19 times has the pre-tourney AdjEM leader won March Madness (2008 Kansas, 2012 Kentucky, 2019 Virginia), and 10 of 19 champions have been a top 3 pre-tourney AdjEM team. 2 teams have been sub-20 in AdjEM: 2003 Syracuse at 19.96 and 2014 UConn at 19.12.
In case anyone’s interested, the average final AdjEM for a tournament winner is about 30.45, compared to 28.20 pre-tourney, a jump of 2.25.
Not saying AdjEM isn’t a great tool for predicting champions, it absolutely is, but there’s missing context behind these numbers.