I have posted 2021 Season: 9.19.21 Simulated RPI Ratings and NCAA Tournament At Large Selections and Seeds. It shows simulated ratings and Tournament at large selections and seeds based on the actual results of games played through Sunday, September 19, and simulated results of games not yet played. Starting with this report, the simulated results of games not yet played are based on current team actual RPI ratings. At this stage of the season, the current RPI ratings and thus the simulated end-of-season results still are primitive, but with each weekly report they will come closer to where they actually will end up. As mentioned in the linked post, on average teams are playing fewer out-of-region games than is normal. There ordinarily are not enough out-of-region games for the RPI properly to rate teams from different regions in relation to each other. This year, that problem will be even worse. This particularly is a problem for teams from the west, which the RPI on average underrates even in a normal year (i.e., they perform better in out-of-region games than their RPI ratings say they should).
I have posted 2021 RPI: 9.19.26 ACTUAL RPI RATINGS, SIMULATED END OF REGULAR SEASON RPI RATINGS, NCAA TOURNAMENT AT LARGE SELECTION AND SEED RANGES, AND AND NCAA TOURNAMENT AT LARGE SELECTIONS AND SEEDS The linked blog article contains two reports: (1) A report showing actual current RPIs and related information based on games played through Sunday, September 26. Among other things, it shows the current ranges of bubble teams for NCAA Tournament at large selections and seeds, based on past history. (2) A report showing simulated ratings and Tournament at large selections and seeds based on the actual results of games played through Sunday, September 26, and simulated results of games not yet played. The simulated results of games not yet played are based on current team actual RPI ratings. At this stage of the season, the current RPI ratings and simulated end-of-season results remain primitive, but with each weekly report they will come closer to where they actually will end up.
I have posted SIMULATED NCAA 2021 TOURNAMENT BRACKET, ALTERNATE METHOD: 9.26.21 It is based on a more complex, but likely more accurate, system for simulating what the NCAA Tournament bracket will be. The linked post includes an explanation of the system. In future weeks, I will post alternative simulated brackets, one using the simpler system I have used so far this year and the other using this more complex system.
I have posted 2021 RPI: 10.3.26 ACTUAL RPI RATINGS (ACTUAL CURRENT AND SIMULATED END OF SEASON), AND SIMULATED NCAA TOURNAMENT AT LARGE SELECTION AND SEEDS This includes my regular weekly actual current RPI report (with historic seed and at large selection candidate ranges as of this stage of the season), my regular weekly simulated end-of-season RPI report, and two simulated NCAA Tournament brackets with one based on my simple system and the other on my more-complex system.
I sometimes get asked where there are good data and similar resources for Division I women’s soccer. Since there are some really good resources out there, I have written an article on my blog that describes and provides links to what I believe are the best resources: ON LINE RESOURCES: DATA, RATINGS, AND BRACKETOLOGY
I have posted 2021 RPI: 10.10.26 RPI RATINGS (ACTUAL CURRENT AND SIMULATED END OF SEASON), AND SIMULATED NCAA TOURNAMENT AT LARGE SELECTIONS AND SEEDS This includes my regular weekly actual current RPI report (with historic seed and at large selection candidate ranges as of this stage of the season), my regular weekly simulated end-of-season RPI report, and two simulated NCAA Tournament brackets with one based on my simple system and the other on my more-complex system.
Penn State had a horrible stretch of losses when their best offensive player (for the season) and 2 defensive players (1 came back, the other is expected back soon) went down at the beginning of conference play. I see they still have a pretty good RPI and strength of schedule. What are their chances of making the NCAA tourney? How many more matches can they afford to lose?
I am not up to answering the last question, but if you look at the most recently linked article, they look like they should get into the Tournament and even presently are in the range as a potential seed (other than a #1 seed).
Barring a collapse of epic proportions and with no tournament, the Ivy race is over. Brown beating Princeton and Harvard in consecutive weeks is impressive. Now the question becomes is has the Ivy done enough non-league to become a 2 or 3 bid league. Harvard and Princeton play next week which will be a huge game on the at-large front. I have to imagine if Harvard runs the table from here they get in.
I have posted 2021 RPI: 10.17.26 RPI RATINGS (ACTUAL CURRENT AND SIMULATED END OF SEASON), AND SIMULATED NCAA TOURNAMENT AT LARGE SELECTIONS AND SEEDS This includes my regular weekly actual current RPI report (with historic seed and at large selection candidate ranges as of this stage of the season), my regular weekly simulated end-of-season RPI report, and two simulated NCAA Tournament brackets with one based on my simple system and the other on my more-complex system.
The NCRPI is the Non-Conference RPI. The NCAA computes it for a team by deleting all of the team’s non-conference games, leaving in all games for all other teams, and then generating a rating using the same formula as for the RPI. The Adjusted NCRPI then takes the NCRPI and applies the same bonuses and penalties as it does for the RPI, but only for the non-conference games. (The bonuses and penalties are pretty inconsequential.) The NCRPI, because based on far fewer games than the RPI, is less reliable as a way of ranking teams. The NCAA believes, however, that it may be better at ranking conferences in relation to each other. So far as DI women’s soccer is concerned, there is a little, but not much, difference in how the ARPI, as compared to the ANCRPI, ranks the conferences in relation to each other. So far as Penn State having the #2 ANCRPI is concerned, it means that Penn State has had an excellent non-conference record against teams that themselves have had very good records. This does not necessarily mean that its non-conference opponents have been good, only that they have had good records against the teams they have played. On looking at their non-conference schedule, however, it looks like on average Penn State played a bunch of pretty good opponents. The ANCRPI is one of the factors the Committee considers in at large selections and seeds for the NCAA Tournament.
I have posted 2021 RPI: 10.24.26 RPI RATINGS (ACTUAL CURRENT AND SIMULATED END OF SEASON), AND SIMULATED NCAA TOURNAMENT AT LARGE SELECTIONS AND SEEDS This includes my regular weekly actual current RPI report (with historic seed and at large selection candidate ranges as of this stage of the season), my regular weekly simulated end-of-season RPI report, and two simulated NCAA Tournament brackets with one based on my simple system and the other on my more-complex system.
Sorry, I just saw this. It is a little more work than I can bite off right now to see about their winning the Tournament, but my current simulation has them as runner up and finishing at #12 in the RPI. A guess is that they might move up to around #8 if they win it.
cpthomas - Enjoy the work, makes for great discussions. 2 questions out of curiosity: 1. On your second list, Harvard is nowhere to be found. Is there something about how the 2nd is done that they don't get a consideration or is it just an omission? 2. Why is BC even on the list? Don't you have to be .500 to make the tourney? You may not know (or care for the sake of this discussion) that they only have 1 game left which is fair, but they can't get to .500.
Love the work/analysis, too--many thanks. Based on nothing more than a review of schedule/results/eye test, I'm not at all seeing Auburn as a 2 seed or SMU as a 3 seed.
Thank you very much for posting! That darned programmer (me) made an error in preparing the simple system report. I now have linked the full proper report to the blog article and also have replaced the table in the article with the correct table. Harvard now is on the list and indeed seeded. At the bottom of the list are a number of teams, simply the ones that are not at large selections but that are next in RPI order. It includes some teams my system projects as falling below the 0.500 winning percentage requirement, including Boston College.
You definitely could be right on Auburn and SMU. Except for the #1 seeds, it is very hard to find a really consistent pattern for the specific seeds the Committee will give other teams. It is a little easier to find a pattern on which teams the Committee will seed, it is just the exact 2, 3, and 4 positions that are hard to get right. In the more complex system seeding, where I have put those teams is based on the factors that do the best job of matching the Committee decisions historically, but that does not mean a whole lot because no factor is that good at matching the Committee decisions. Using the more complex system for #2 seeds, TCU was the only clear #2. If Auburn is out as a #2, next in line would be Georgetown. And for #3 seeds, Southern California was the only clear #3. If SMU is out as a #3, next in line would be one of Tennessee, Princeton, or Michigan. Also, the more complex system table only shows three #4 seeds. The fourth should be Princeton (but could be Pepperdine, Notre Dame, UCLA, Michigan, or South Florida). Hopefully, some of this will become clearer over the next week and a half.
Seeding carnage as Georgetown, Auburn, Ole Miss and SMU all lost last night--with SMU getting beat by Memphis 5-0. If you get beat 5-0 in your last league game, in a less than stellar conference, you perhaps don't deserve a national seed--though, yes, body of work is the most important determinant.
I think if Southern Cal or UCLA wins the PAC-12 then they have to be a #2 seed, right? Would UCLA have a shout at a #1 if they go unbeaten?
For the life of me I do not understand why Brown and Harvard are rated so highly. Princeton at least has results against Rutgers and Georgetown. But Harvard has zero Top 50 wins. Brown’s only Top 50 wins are Harvard and Princeton. And they might actually get a seed? Madness. Also Brown’s RPI ranking rose 7 spots inside the top 20 to #13 after beating a terrible Penn team? How?? Why???
The terrible Penn team is 9-4-2 so they get a boost. That’s how the RPI works. It doesn’t have opinions. It just goes of your record, your opponents’ records, and the record of the opponents’ opponents. Penn has played