When the Basketball Power Index (BPI) was introduced last week as a method to rate college basketball teams, it was touted as a superior method of gauging a team's true merit than the RPI -- the metric to rank teams used by the NCAA selection committee each March. The Sagarin and KenPom ratings have existed as good alternatives to the RPI for years, and BPI is another step in the evolution of these more advanced metrics. But how do we know that BPI is a better representation of a team's overall performance than RPI?
First off, the system is more complete and mathematically sound than the RPI. BPI accounts for not only who won and lost, but by how much, in how many possessions, and whether either team was missing key players. Even with elements that both systems capture -- like strength of schedule -- the BPI's way of doing so is much more thorough and consistent.
But that's all just theory. You probably want to know if the BPI does a better job in predicting tournament games than the RPI does, and if so, how much better.