Annually I raise the alarm about our how poorly planned our basketball schedule is. My primary complaint is this: the Selection Committee relies heavily on the RPI, yet our schedule isn’t designed to take advantage of this easily manipulated metric. It doesn’t matter that the RPI is a ridiculous model that has about as much predictive power as a coin flip, what matters is that it dominates the data the committee uses to select the best at-large teams for the Tourney. Smart scheduling can easily make 2 seeds worth of difference come March, and obviously be the difference between the NCAA’s and going to the NIT. So was I barking up the wrong tree this year, or is our schedule really going to cost us come Selection Sunday?
Using a replica of the NCAA Selection Sheet, which is the primary source of information the committee members use, I looked at our schedule utilizing the RPI and then compared that to the premier statistical model developed by Ken Pomeroy. No stats based model is going to perfectly predict how strong a team is, but Pomeroy’s is the best. If you want to see the value of the RPI then start betting the money line in Vegas based on RPI’s and then give me a call when you’re pawning your valuables to pay the bill.
This table gives the key indicators based on RPI and KenPom:
**edit - I'm checking with the site to see how non-conference SOS was calculated, as that's the highest I've seen**
|Strength of Schedule||82||54|
What would this mean for our seeding? Tough to say without looking at every other team, but our resume would certainly be more impressive than the one the committee will be examining next Sunday.