Every once in a while, as a byproduct of a project I'm working on, I run into something interesting. I'm currently working on a revision and update of my annual review of the Women's Soccer Committee's patterns when doing NCAA Tournament seeds and at large selections. In the course of that work, I noticed a truly interesting bit of information:
Suppose the Committee simply used the NCAA RPI to make at large selections. To do that, it would have put the teams in NCAA RPI rank order and selected the top 33 or 34 teams (depending on the year and how many conferences there were that year), that were not Automatic Qualifiers, to fill the at large positions.
Each data point in the chart shows, for the particular year, the percent of the Committee's actual at large selections that the NCAA would have matched if it simply had used teams' NCAA RPI ranks as the basis for selection. For example, looking at 2007, the Committee actually selected 34 at large teams. If the NCAA simply had used the NCAA RPI as the basis for selection, it would have matched 32 of those teams, which is 94.1% of them. This 94.1% match is an indicator of how much the NCAA RPI influenced the Committee.
Across the chart from left to right, you can see a chronological picture of how well the NCAA RPI rankings matched the Committee's actual decisions. The straight black line is a computer generated straight trend line showing the trend in the Committee's decisions. While the percentages go up and down, the chart suggests that the NCAA RPI, over recent years, has had a decrease in influence over the Committee's decisions. It's not a big decrease, but it's noticeable.
The apparent decrease could be only a random variation and not really signify a decrease in influence. On the other hand, it is consistent with the Committee's recently having asked for and gained the ability to supplement the NCAA RPI with a different rating system, which itself suggests waning Committee trust of the NCAA RPI.
No comments:
Post a Comment