|Posted by Lobofan2003 on January 30, 2019 at 3:35 AM|
Now that we are near the end of January, in the middle of conference play, this is the point in the season where you called be more confident in using the RPI as a metric as a bracketologist and that there wouldn't be wild swings in the RPI rankings every night. Yes, the RPI was a rankings system with obvious flaws for simplicity, but they were at least a way to get a relative idea to get at the relative strength of a team.
Now that the NCAA NET rankings are being used, how does a team's NET ranking vs. their RPI ranking compare at this point in the season and are there any obvious bugs or flaws in the NET rankings?
Remember that the formula for NET includes things like scoring margin (capped at 10 per game), offensive/defensive efficiency rankings, quality of opponents played, road wins, etc. (Reminder that the RPI did not control for scoring margin and efficiency rankings but did attempt to control for Strength of Schedule and additional weighting to road wins)
As such, here are the Top teams in NET and RPI as of 1/29/19 from the NCAA website and realtimerpi.com
From the above rankings you can see that three teams (Nebraska, NC State and Florida) have seen a large bump in their NET rankings vs. their corresponding RPI rankings. Other college basketball analytics gurus have surmised, by looking at efficiency rankings, that the reason for many of these bumps is from an increase in offensive efficiency rankings from blow outs against teams that would negatively bring down the team's overall strength of schedule. NC State is the most glaring example of this, their SOS is 352 but they have a very strong NET ranking based on their efficieny rankings.
The fact that offensive efficiency can be influenced by scoring a lot of points in blowouts against lesser competition points to what some critics call a flaw in the rankings that allows the NET to be "manipulated" by running up the score against lower competition. To be sure, critics of the RPI said that it could be manipulated as well by scheduling lots of teams with middle of the road competition on the road/in neutral locations and encouraging others in your conference to do that as well.
As the first year of the NET rankings I am sure that adjustments will be made in future years to address some of these concerns. Time will tell whether or not teams that would not normally have been considered due to their SOS/RPI are now considered based on their NET rankings and how much it will influence the Committee. Keep in mind that the NET is only one ranking system that Committee members can use, they also have access to KenPom, ESPN's BPI, RPI and other ranking systems so I am hopeful that the Committee will be balanced in their selection approach this year.
One thing is for certain, bracketologists are in uncharted territory this year!