|
Methodology of Scott Gruhn's Computerized Rating System:
Strength of Schedule: The NCAA's RPI does not compute how good the opposing team is. It merely figures in win / loss percentages, which does not acceptably measure the quality of a team. A team in a weak conference or with a weak schedule will have a better percentage, which skews the result. Scott Gruhn's rating system computes how good both teams are, and then, based on the margin of victory, determines how the game affected the estimated quality of each team. The ratings implications from each game are averaged for each team. Now with this better gauge of the quality of each team, the process goes through a second iteration to more precisely determine the quality of all Division I teams, then a third iteration, and so on, until the results are as refined as possible.
Margin of Victory: The NCAA's RPI does not consider whether your team barely scraped past the opposing team or whether it blew them away. It only considers whether your team won or lost. Because of this, the RPI is incapable of adequately comparing two teams that both beat the same team but by differing margins. This simulation, on the other hand, is based on the margin of victory or defeat, as expressed as a ratio of scores.
Home or Away Games: In college basketball, there is a home-court advantage because of the crowd cheering and pressuring the visiting team. This home-team advantage, amounting to approximately 5% of the final score, is considered by this simulation but not by the NCAA's RPI.
Time Relevance of Game: The college basketball season is nearly six months long. Some teams, young teams in particular, take most of the season before they are playing their best game. It takes time to get into game shape, and it takes time to learn the habits of teammates. The NCAA considers a game four months ago as valid an indicator as a game played yesterday, which inaccurately estimates the team's current strength. In Scott Gruhn's simulation, the final game of the season is considered nearly twice as good an indicator of the team's final strength as the first game of the season.
|
|
