FWIW
- Thread starter ScarletteK80
- Start date
RPI is an imperfect algorithm, similar to the Sagarin ratings in College Football. As the season goes on and there is more data to compute, the RPI improves somewhat. It is just one of many factors that the NCAA committee will look at to determine, in their judgment , who the 64 teams that go into the tournament are, and where they should be seeded.
What Abro says, sort of. The more data, the better the results, although, like most systems, there are flaws in the results. RPI can be gamed, at least somewhat, by careful selection of who you play; Rutgers certainly did that at times in the past. All of that is no particular secret, articles have been written about how to influence your RPI.
SOS is 75% of the RPI, so it goes a long way to determining where you rank. However, your record, while only 25%, has the greatest influence per single event (for example winning) - it has by far the smallest quantity of events and is, FWIW, most within your own control.
In practical terms, the part of RPI that has the most impact is W/L record because it can vary from 0.000 to 1.000. SoS has a much narrower range, so even though it's 75%, it probably has much less impact than W/L. (And SOS actually has two components - opponents' records and opponents' opponents' records.) SOS also steadily becomes less important as a factor because the conference season drags everyone's SOS towards the middle of the range - the records of your conference opponents against each other are going to end of close to .500.