At the time of the latter post, I made a mental note to come back to it after the season, as it gives me a chance to explain a valid use of the RPI before the end of the season and a not valid use. First, the RPI is not a predictive tool. Rather, to use a term I learned from someone else, it is a retrodictive tool. What that that means is, it looks backwards and tells how teams have done, not forward telling how they will do. There is a sense in which you can use it as a predictive tool, however. With a supplemental aid, starting at least at the end of week 6 of the season it is good at predicting pools of teams from which will come #1, 2, 3, and 4 seeds and at large selections to the NCAA tournament and which teams will not get seeds and at large selections. I will use Florida State as an example. Florida State, at the end of week 6 of the regular season, had an RPI rank of #14. Based on past history, as of that week of the season any team ranked #28 or better is a potential #1 seed. Any team ranked #29 or poorer will not get a #1 seed. At the end of week 7, the rank was #16. Again, the pool of potential #1 seeds is #28 or better. At the end of week 8, the rank was #9 and the pool was limited to #19 or better. At the end of week 9, the rank was #5 and the pool again was #19 or better. At the end of week 10, the rank was #5 and the pool was #14 or better. At the end of week 11, the rank was #4 and the pool was #10 or better. At the end of week 12 (end of season), the rank was #1 and the pool was #7 or better. Thus starting about half way through the season, we knew from the RPI that Florida State was a potential #1 seed. Further, we safely could predict that teams ranked outside the above pool limits would not get #1 seeds. There are similar history based pool limits for #2, 3, and 4 seeds and for at large selections. Thus although early season RPI ranks are not good predictors for what the ultimate ranks will be, they are good indicators of which teams have the potential for NCAA Tournament seeds and at large selections and are good predictors of which teams will not get seeds and at large selections. Thus the early season RPI ranks can be useful, if one knows how to use them properly.
You can find my annual discussion of interesting decisions the Committee made in forming the 2022 NCAA Tournament bracket here: REVIEW OF DECISIONS THE NCAA D1 WOMEN’S SOCCER COMMITTEE MADE FOR THE 2022 NCAA TOURNAMENT Two things stand out to me this year: 1. The Committee appears to have assigned less weight to poor tie results than historically has been the case. This is not surprising since the change to no overtimes during the regular season resulted in there being twice as many ties this year as there have been in the past. The doubling of the number of ties is exactly as expected since if all the games decided in the past by golden goal would have been ties, it would have doubled the number of tie games. 2. The Committee appears to have given less weight to the strength of a team’s conference -- both as a positive factor for teams from strong conferences and as a negative factor for teams from weak conferences -- than might have been expected based on past Committee decisions. This does not necessarily mean the Committee has changed the weight it gives to conference strength. It may simply mean that the Committee had team profiles it has not had in the past and these forced it to make choices it has not had to make before about how much weight to give conference strength.