The Big Ten is rated as the best conference by the Ratings Percentage Index (RPI), a primary tool used by the selection committee. That is not unique, and I would say that every objective ranking system I have seen has the Big Ten at #1 based on the results this year. Ken Pomeroy’s and Jeff Sagarin’s ratings both have the Big Ten with 4 of the top 9 teams, and 8 of their 12 teams inside the top 50.
Consensus would seem to pretty solidly put the Big Ten at #1 looking at the depth and average rating of the conference. So why do I think they are probably overrated? Because almost every team rated as the “best conference” is overvalued based on a relatively small sample size of results, most of which occurred in November and December.
Before I get into some numbers, first some philosophy. The top ranked conference would almost always be overrated because there is no place to go but down. Now imagine a hypothetical where we have ten teams, and we clone them. We know they are exactly equal. Still, from game to game, we have random variation from things like missed shots, foul trouble, temporary injuries, or just the inconsistency of players age 18-23. What is the likelihood that if we simulate results, we actually end up measuring all of our cloned teams as being equal? I would think pretty slim. Some would appear better than others.
Now, extrapolate that across conferences, and a handful of games against similar competition. A few bad games by one conference and good ones by another could cause us to view one as better than the other, when it may not be the underlying truth.
So, those ideas led me to look at how conferences have actually ranked by the RPI entering the tournament for the last decade, and comparing how they performed relative to expectations. Performance Above Seed Expectation (PASE) allows us to compare a team to all others who had the same tournament seed by average wins, since we know that #2 seeds win more than #12 seeds. But by using PASE we can see if a team, or in this case, a conference, over performs or underperforms the average expectation based on how RPI ranked that conference.
I divided the conferences into the six power conferences (ACC, Big East, Big Ten, Big XII, Pac-10/12, and SEC) and also found the six highest ranked other conferences each year–usually a combination of the Mountain West, Atlantic-10, Missouri Valley, and Conference USA, plus a mix of the WAC, WCC, Horizon, CAA and MAC. The identity of the top rated power conference changes from year to year, and every conference except the Pac-10/12 has been #1 at least one year over the last decade. First, for the power conferences, here is the PASE results based on conference rank by RPI entering the tournament.
The best conference by RPI has, in fact, underperformed their win total by almost 11 wins over the last decade. Half of the “top” conferences have underperformed seed expectation by at least 2 wins, including the Big East last year (despite UCONN winning as a #3 seed). We see an inverse relationship between RPI rank of the conference entering the tournament, and whether the conference tends to underperform or over perform based on its seed.
Here’s the same summary for conference rank among the other conferences:
Similar pattern, with the two highest ranked “other” conferences underperforming their seeds collectively, while the rest do well. Among the power conferences, the two highest ranked by RPI underperformed seed by 16.8 wins, while the rest over performed by 14.4 wins. Among the other conferences, the top two underperformed by 12.1 wins, and the next four over performed by 11.4 wins.
What does this mean for the Big Ten? I don’t really think the Pac-12 and Big Ten are equal. What I do think is that the magnitude of the difference is exaggerated by the results. Perhaps the Big Ten can overcome, just like the SEC did in 2007 (when Florida won the title from the #1 RPI conference, and SEC was +4.0 wins better than expected) or the ACC did in 2004 (when Duke and Georgia Tech reached the Final Four, +2.8 wins better than expected).
The last time the Big Ten was #1 in RPI, though, It wasn’t pretty. In 2006, Iowa lost in the first round as a #3 seed, Ohio State got knocked off by Georgetown in the second round as a #2 seed, Michigan State had a rare first round loss to tournament surprise George Mason, and the Big Ten went 3-6 and got no teams to the Sweet 16 despite having five teams seeded #6 or higher.
More likely, they are a little overvalued because they played well in November and December. If history is an indicator, they are going to get about 7 teams in. However, you might look at SEC and ACC teams as being undervalued in the tournament among the mid to lower seeds, and the few teams from the Pac-12 who get in may be under seeded based on those early results pulling them down.
[photo via Getty]
blog comments powered by Disqus